From 00f4c64c80acd83662932b83aa3ec16aa2620b17 Mon Sep 17 00:00:00 2001 From: linmoumou Date: Mon, 27 Apr 2026 00:19:36 +0800 Subject: [PATCH 1/9] docs: add partner-node tutorials for ElevenLabs, Grok, Vidu, MiniMax, Magnific, PixVerse, Quiver, Veo, Topaz, HitPaw, WaveSpeed Each provider gets a dedicated mdx page with overview, key capabilities, and available workflows (Cloud Run + Download JSON cards). Follows the existing Seedance 2.0 page format. --- .../partner-nodes/elevenlabs/elevenlabs.mdx | 102 +++++++++++++ tutorials/partner-nodes/grok/grok.mdx | 102 +++++++++++++ tutorials/partner-nodes/hitpaw/hitpaw.mdx | 50 +++++++ tutorials/partner-nodes/magnific/magnific.mdx | 90 ++++++++++++ tutorials/partner-nodes/minimax/minimax.mdx | 67 +++++++++ tutorials/partner-nodes/pixverse/pixverse.mdx | 68 +++++++++ tutorials/partner-nodes/quiver/quiver.mdx | 59 ++++++++ tutorials/partner-nodes/topaz/topaz.mdx | 69 +++++++++ tutorials/partner-nodes/veo/veo.mdx | 57 ++++++++ tutorials/partner-nodes/vidu/vidu.mdx | 135 ++++++++++++++++++ .../partner-nodes/wavespeed/wavespeed.mdx | 59 ++++++++ 11 files changed, 858 insertions(+) create mode 100644 tutorials/partner-nodes/elevenlabs/elevenlabs.mdx create mode 100644 tutorials/partner-nodes/grok/grok.mdx create mode 100644 tutorials/partner-nodes/hitpaw/hitpaw.mdx create mode 100644 tutorials/partner-nodes/magnific/magnific.mdx create mode 100644 tutorials/partner-nodes/minimax/minimax.mdx create mode 100644 tutorials/partner-nodes/pixverse/pixverse.mdx create mode 100644 tutorials/partner-nodes/quiver/quiver.mdx create mode 100644 tutorials/partner-nodes/topaz/topaz.mdx create mode 100644 tutorials/partner-nodes/veo/veo.mdx create mode 100644 tutorials/partner-nodes/vidu/vidu.mdx create mode 100644 tutorials/partner-nodes/wavespeed/wavespeed.mdx diff --git a/tutorials/partner-nodes/elevenlabs/elevenlabs.mdx b/tutorials/partner-nodes/elevenlabs/elevenlabs.mdx new file mode 100644 index 000000000..a69de824b --- /dev/null +++ b/tutorials/partner-nodes/elevenlabs/elevenlabs.mdx @@ -0,0 +1,102 @@ +--- +title: "ElevenLabs Text-to-Speech, Speech-to-Text, Sound Effects & Voice Tools" +description: "Generate natural, emotionally rich speech, transcribe audio, create sound effects, and isolate voices using ElevenLabs in ComfyUI." +sidebarTitle: "ElevenLabs" +--- + +import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; +import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; + +ElevenLabs brings world-class voice AI to ComfyUI, covering the full spectrum of audio generation and processing. **Eleven v3** delivers the most emotionally expressive speech, capturing subtle vocal nuances in intonation, pacing, and emphasis for character voiceovers and dramatic narration. **Eleven Multilingual v2** excels at stable, consistent long-form output across 29 languages, making it the choice for audiobooks and localized content. **Eleven Flash v2.5** provides near-instant speech generation with ~75ms latency at roughly half the cost, perfect for live streaming, chatbots, and real-time applications. **Scribe v2** transcribes audio to text with support for 90+ languages, speaker diarization for up to 32 speakers, 56-class entity detection, keyword prompting, and precise word-level timestamps. Beyond speech, ElevenLabs also generates sound effects from text descriptions, isolates voice from background noise, creates multi-character dialogue, and transforms one voice into another while preserving the original delivery. + +## Key capabilities + +- **Text to Speech** — Generate lifelike audio from text with three model tiers (Eleven v3, Multilingual v2, Flash v2.5) +- **Speech to Speech** — Transform a source audio clip into a different voice while keeping the original intonation and pacing +- **Speech to Text** — Transcribe audio with Scribe v2: 90+ languages, 32-speaker diarization, entity detection, keyword boost +- **Sound Effects** — Generate audio effects from text descriptions +- **Voice Isolation** — Separate voice from background noise, music, or ambient sound +- **Multi-Character Dialogue** — Generate conversations between multiple distinct voices in a single pass + + + + +## Available workflows + +### Text to Speech + +Generate speech from text using the flagship Eleven v3, the stable Multilingual v2, or the low-latency Flash v2.5 model. + + + + Generate speech from text directly in your browser. + + + Download the workflow JSON. + + + +### Speech to Speech + +Take an existing audio recording and transform it into a different voice or style while preserving the spoken content and delivery. + + + + Transform existing audio into a new voice. + + + Download the workflow JSON. + + + +### Speech to Text + +Transcribe audio to text using Scribe v2, with advanced speaker diarization and entity detection. + + + + Transcribe audio to text directly in your workflow. + + + Download the workflow JSON. + + + +### Text to Sound Effects + +Generate sound effects, ambient audio, and foley from text descriptions. + + + + Generate sound effects from text. + + + Download the workflow JSON. + + + +### Voice Isolation + +Separate the voice track from background noise, music, or ambient sound in any audio recording. + + + + Isolate voice from noisy audio. + + + Download the workflow JSON. + + + +### Text to Dialogue + +Generate multi-character dialogue with distinct voices from a single text input. + + + + Generate multi-voice dialogue from text. + + + Download the workflow JSON. + + diff --git a/tutorials/partner-nodes/grok/grok.mdx b/tutorials/partner-nodes/grok/grok.mdx new file mode 100644 index 000000000..ce893de7f --- /dev/null +++ b/tutorials/partner-nodes/grok/grok.mdx @@ -0,0 +1,102 @@ +--- +title: "Grok Image & Video Generation" +description: "Generate images and videos with xAI's Grok models — fast, moody aesthetics with strong anime and cinematic character rendering." +sidebarTitle: "Grok" +--- + +import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; +import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; + +Grok Imagine by xAI brings a distinct visual style to AI image and video generation. Its **grok-imagine-image-pro** model produces the highest quality output with a signature "moody" aesthetic — dramatic lighting, filmic contrast, rich color saturation that makes every frame feel cinematic. The **grok-imagine-image** standard tier offers faster generation at ~4 seconds per image while maintaining strong visual quality, and **grok-imagine-image-beta** provides early access to experimental features. Grok excels at anime and cyberpunk aesthetics, cinematic character renders, 2D illustrations, and dynamic graphical compositions. For video, Grok supports text-to-video, video editing, video extending, and reference-to-video generation with up to 7 reference images for consistent character and scene identity across clips. + +## Key capabilities + +- **Text to Image** — Generate images from text prompts with pro/standard/beta quality tiers +- **Image Editing** — Edit existing images with text instructions +- **Text to Video** — Generate video from text descriptions +- **Video Editing** — Edit existing video footage +- **Video Extending** — Extend an input video (15s in, extend up to 10s) +- **Reference to Video** — Generate consistent video using up to 7 reference images + + + + +## Available workflows + +### Text to Image + +Generate images from text prompts using the highest quality model for pro-tier results. + + + + Generate images from text instantly. + + + Download the workflow JSON. + + + +### Image Editing + +Edit existing images using text-based instructions for modifications. + + + + Edit images with text instructions. + + + Download the workflow JSON. + + + +### Text to Video + +Generate video from text descriptions with Grok's signature visual style. + + + + Generate video from text. + + + Download the workflow JSON. + + + +### Video Editing + +Edit existing video footage using text instructions. + + + + Edit existing video with text. + + + Download the workflow JSON. + + + +### Video Extending + +Extend an existing video clip by up to 10 seconds, creating natural continuations rather than loops. + + + + Extend a video clip. + + + Download the workflow JSON. + + + +### Reference to Video + +Generate consistent video content guided by up to 7 reference images for identity, scene, and narrative consistency. + + + + Generate video with reference images. + + + Download the workflow JSON. + + diff --git a/tutorials/partner-nodes/hitpaw/hitpaw.mdx b/tutorials/partner-nodes/hitpaw/hitpaw.mdx new file mode 100644 index 000000000..3a396f657 --- /dev/null +++ b/tutorials/partner-nodes/hitpaw/hitpaw.mdx @@ -0,0 +1,50 @@ +--- +title: "Image & Video Enhancement with HitPaw" +description: "Enhance and upscale images and video using HitPaw's AI models within ComfyUI — covering general image enhancement and video enhancement workflows." +sidebarTitle: "HitPaw" +--- + +import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; +import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; + +HitPaw is a well-established name in AI-powered media enhancement, known for its accessible, high-quality image and video upscaling tools. Now available as native nodes within ComfyUI, HitPaw brings its production-grade enhancement capabilities directly into your visual workflow, eliminating the need to switch between applications for post-processing. + +The ComfyUI integration covers two core workflows: **general image enhancement** (utility_hitpaw_general_image_enhance) and **video enhancement** (utility_hitpaw_video_enhance). Both are powered by HitPaw's proprietary AI models running on ComfyCloud infrastructure, meaning no local GPU resources are consumed during processing. + +The HitPaw general image enhancement model is designed as a versatile all-rounder. It handles common image quality issues including low resolution, noise, blur, and compression artifacts. Whether you're upscaling AI-generated images from Stable Diffusion or Midjourney, restoring old photographs, or preparing product images for e-commerce, HitPaw produces clean, natural-looking results without the over-sharpened or "plastic" appearance that some upscalers introduce. The model intelligently distinguishes between different types of image content — portraits, landscapes, architecture, text — and applies enhancement strategies optimized for each category. + +On the video side, HitPaw video enhancement performs frame-by-frame AI upscaling and denoising. It is particularly effective for improving the quality of compressed video footage, screen recordings, and low-resolution source material. The model maintains temporal consistency across frames, preventing the flickering or "pumping" artifacts that can occur with naive frame-independent processing. This makes it suitable for everything from archival video restoration to improving the quality of AI-generated video clips before final export. + +Both workflows are designed to be straightforward drop-in enhancements. You can connect them to any image or video output node in ComfyUI — whether from generative models, renders, or imported media — and get consistently good results without extensive parameter tuning. The HitPaw nodes are an excellent choice for creators who want reliable, predictable enhancement quality with minimal configuration overhead. + +## Key capabilities + +- **General image enhancement** — Multi-purpose AI upscaling and restoration that handles noise, blur, artifacts, and low resolution across photographs, renders, and AI-generated images. +- **Video enhancement** — Frame-by-frame video upscaling and denoising with temporal consistency, suitable for compressed footage, screen recordings, and archival video. +- **Content-aware processing** — Automatically detects image content type (portrait, landscape, text, etc.) and applies optimized enhancement strategies for natural results. +- **No local GPU required** — All processing runs on ComfyCloud's infrastructure, freeing your local hardware for other tasks. +- **Minimal configuration** — Simple drop-in nodes that produce reliable results without extensive parameter tuning, ideal for integration into larger ComfyUI pipelines. +- **Broad format support** — Compatible with common image and video formats, making it a versatile post-processing step for any ComfyUI workflow. + + + + +## Available workflows + +### General Image Enhancement + +Upscale and restore images with HitPaw's all-purpose AI enhancement model. Handles low resolution, noise, blur, and compression artifacts in a single pass. + + + Launch the image enhancement workflow on ComfyCloud + Download the workflow JSON for local ComfyUI + + +### Video Enhancement + +Enhance and upscale video footage frame-by-frame using HitPaw's video AI. Improves resolution, reduces noise, and maintains temporal smoothness across frames. + + + Launch the video enhancement workflow on ComfyCloud + Download the workflow JSON for local ComfyUI + diff --git a/tutorials/partner-nodes/magnific/magnific.mdx b/tutorials/partner-nodes/magnific/magnific.mdx new file mode 100644 index 000000000..8343bb839 --- /dev/null +++ b/tutorials/partner-nodes/magnific/magnific.mdx @@ -0,0 +1,90 @@ +--- +title: "Magnific AI API Node Image Enhancement ComfyUI Official Examples" +description: "This guide covers how to use the Magnific partner nodes in ComfyUI for AI-powered image enhancement, upscaling, style transfer, relighting, and skin enhancement." +sidebarTitle: "Magnific AI" +--- + +import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; +import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; + +Magnific AI is a cutting-edge AI image enhancement platform that specializes in high-quality upscaling, creative image transformation, and intelligent photo retouching. Now natively integrated as partner nodes in ComfyUI, Magnific brings professional-grade image enhancement capabilities directly into your workflows. + +The Magnific node suite offers five distinct modes to handle a wide range of image enhancement tasks. **Precise Upscale** delivers pixel-perfect enlargement with minimal deviation from the original, making it ideal for preserving fidelity in product shots or archival photos. **Creative Upscale**, in contrast, uses AI-driven semantic interpretation to add realistic detail at up to 16x amplification — perfect for breathing new life into low-resolution images. The **Creativity slider** in Creative mode gives you fine-grained control over how much new detail the model introduces, from subtle texture enhancement to bold reinterpretation. **Style Transfer** applies a reference image's aesthetic — color palette, lighting mood, brushstroke texture — onto your source image, enabling artistic transformations without manual masking. **Relight** lets you control the lighting direction, intensity, and color temperature, making it easy to match subject lighting with a target environment. **Skin Enhancer** provides intelligent portrait retouching by smoothing skin texture while preserving natural features, reducing blemishes and evening out skin tone automatically. + + + + +## Key capabilities + +- **Precise Upscale** — High-fidelity image enlargement that preserves original details and introduces minimal new content; ideal for archival or product images. +- **Creative Upscale (16x)** — AI-powered upscaling that interprets and adds realistic detail at up to 16x magnification; the Creativity slider controls how much new detail is generated. +- **Style Transfer** — Transfers the artistic style, color palette, and lighting from a reference image onto your source image for creative transformations. +- **Relight** — Adjusts lighting direction, intensity, and color temperature on existing images, enabling seamless compositing into new environments. +- **Skin Enhancer** — Intelligent portrait retouching that smooths skin, reduces blemishes, and evens skin tone while preserving natural facial features. + +## Available workflows + +### Image Upscale — Precise + +Generate a high-fidelity upscaled version of your input image with precise detail preservation. This mode adds minimal new content and stays as close to the original as possible. + + + + Run the precise image upscale workflow instantly on Comfy Cloud. + + + Download the workflow JSON file for local use. + + + +### Image Upscale — Creative + +Generate a creatively upscaled version of your input image at up to 16x magnification. The Creativity slider allows you to control how much new AI-generated detail is added — lower values stay closer to the original, while higher values produce richer, more detailed results. + + + + Run the creative image upscale workflow instantly on Comfy Cloud. + + + Download the workflow JSON file for local use. + + + +### Style Transfer + +Apply the visual style, color palette, and lighting mood from a reference image to your source image. This enables artistic reinterpretations — turn a photograph into an oil painting, or match the aesthetic of a cinematic frame. + + + + Run the style transfer workflow instantly on Comfy Cloud. + + + Download the workflow JSON file for local use. + + + +### Relight + +Adjust the lighting on your image by specifying direction, intensity, and color temperature. This is especially useful for compositing subjects into new scenes where the original lighting doesn't match. + + + + Run the relight workflow instantly on Comfy Cloud. + + + Download the workflow JSON file for local use. + + + +### Skin Enhancer + +Apply intelligent portrait retouching to smooth skin, reduce blemishes, and even out skin tone while preserving natural facial features and texture. + + + + Run the skin enhancer workflow instantly on Comfy Cloud. + + + Download the workflow JSON file for local use. + + diff --git a/tutorials/partner-nodes/minimax/minimax.mdx b/tutorials/partner-nodes/minimax/minimax.mdx new file mode 100644 index 000000000..3730ec3af --- /dev/null +++ b/tutorials/partner-nodes/minimax/minimax.mdx @@ -0,0 +1,67 @@ +--- +title: "MiniMax (Hailuo) - AI video generation" +description: "Generate high-quality video from text, images, and subject references using MiniMax's T2V-01, I2V-01, and S2V-01 models in ComfyUI" +sidebarTitle: "MiniMax (Hailuo)" +--- + +import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; +import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; + +MiniMax, also known as Hailuo, is a leading AI video generation platform recognized for its exceptional motion quality and cinematic output. Now available inside ComfyUI, MiniMax offers three core model families that cover a comprehensive range of video generation use cases — from cinematic narrative production to character-driven content creation. + +T2V-01 is the text-to-video model that supports the Director mode with 15 distinct camera movement types: pan, tilt, dolly, crane, tracking, aerial, handheld, roll, zoom-in, zoom-out, push-in, pull-out, orbit, first-person, and static. This gives you granular control over shot composition and cinematic storytelling without leaving the ComfyUI node graph. I2V-01 takes image-to-video generation further with two sub-modes: I2V-01-Director for stylized camera direction with image input, and I2V-01-Live for generating dynamic, lively motion from a single image, optimized for natural movement like flowing hair, swaying grass, and gentle camera breathing. S2V-01 introduces subject reference to video, allowing you to supply a reference image of a specific character or object and maintain its identity throughout the generated clip — essential for branded content, serial characters, and product consistency. + +All MiniMax models are known for industry-leading motion quality, with realistic physics in character movement, fluid dynamics, and environmental effects. The Director mode is particularly powerful for content creators who need precise camera language to tell their stories — each of the 15 movement types produces a distinctly different cinematic feel, from the intimacy of handheld to the grandeur of aerial crane shots. + +## Key capabilities + +- **T2V-01 with Director mode** — 15 cinematic camera movement types (pan, tilt, dolly, crane, tracking, aerial, handheld, roll, zoom-in, zoom-out, push-in, pull-out, orbit, first-person, static) for precise shot-level direction +- **I2V-01-Director** — Apply cinematic camera controls to image-initiated videos, combining the visual foundation of a starting image with the expressive language of Director mode +- **I2V-01-Live** — Generate natural, lively motion from a single image with organic movement dynamics — ideal for bringing portraits, landscapes, and product shots to life +- **S2V-01 subject reference** — Maintain character, face, or object identity across the generated video using a single reference image; perfect for serial content +- **High motion quality** — Industry-leading realism in character movement, physics, and fluid dynamics across all model variants +- **Flexible input modes** — Text-only, image+text, and subject reference+text, each with optional camera direction + + + + +## Available workflows + +### Text to video (T2V-01) + +Generate a video from a text prompt with full Director mode support. Choose from 15 camera movement types — pan, tilt, dolly, crane, tracking, aerial, handheld, roll, zoom-in, zoom-out, push-in, pull-out, orbit, first-person, and static — to craft the exact cinematic feel you need. + + + + Try the T2V-01 workflow instantly on Comfy Cloud. + + + Download the workflow JSON. + + + +### Image to video (I2V-01) + +Animate a starting image into a video clip. Two modes are available: I2V-01-Director brings camera direction (pan, zoom, tilt, etc.) to image-based generation, while I2V-01-Live prioritizes natural, organic motion — perfect for bringing portraits, landscapes, and product shots to life with subtle movement and atmospheric breathing. + + + + Try the I2V-01 workflow instantly on Comfy Cloud. + + + Download the workflow JSON. + + + +### Subject reference to video (S2V-01) + +Generate a video while maintaining the identity of a specific character, face, or object throughout the clip. Supply a reference image alongside your text prompt, and S2V-01 ensures consistent appearance across the entire generated sequence. This is especially useful for character-driven narratives, branded content, and product showcases where visual identity must remain stable from frame to frame. + + + + Try the S2V-01 workflow instantly on Comfy Cloud. + + + Download the workflow JSON. + + diff --git a/tutorials/partner-nodes/pixverse/pixverse.mdx b/tutorials/partner-nodes/pixverse/pixverse.mdx new file mode 100644 index 000000000..ec26df97b --- /dev/null +++ b/tutorials/partner-nodes/pixverse/pixverse.mdx @@ -0,0 +1,68 @@ +--- +title: "PixVerse AI API Node Video Generation ComfyUI Official Examples" +description: "This guide covers how to use the PixVerse partner nodes in ComfyUI for AI-powered video generation from text prompts, images, and template-based workflows." +sidebarTitle: "PixVerse AI" +--- + +import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; +import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; + +PixVerse is a state-of-the-art AI video generation platform that produces high-quality, cinematic videos from text descriptions and reference images. Now integrated as partner nodes in ComfyUI, PixVerse brings its powerful video generation capabilities directly into your workflows without requiring a separate API key or external setup. + +The PixVerse node suite currently supports three generation modes. **Text-to-Video** allows you to create original videos from nothing but a text prompt — describe a scene, a character, an action, or a cinematic mood, and PixVerse generates a corresponding video clip. This is ideal for concept visualization, storyboarding, and rapid ideation. **Image-to-Video** takes a reference image as input and animates it, preserving the subject and scene composition while introducing motion. This mode excels at bringing static artwork, product shots, or photographs to life with natural movement. **Template Image-to-Video** provides a more structured approach: it uses a pre-defined motion template applied to your input image, giving you predictable and repeatable animation styles without the variability of open-ended generation. This is especially useful for production pipelines where consistent motion characteristics are required across multiple assets. + +All three modes share a common set of parameters for fine-tuning the output, including resolution, duration, and style guidance. The generated videos can be combined with other ComfyUI nodes for further processing — overlay text, composite with other clips, or pipe into a video-to-video enhancement workflow. + +PixVerse is particularly well-suited for content creators, marketers, and artists who need rapid video prototyping without the overhead of traditional animation pipelines. For example, a marketer can use Text-to-Video to generate product demonstration clips from copy alone, while a digital artist can use Image-to-Video to breathe motion into their illustrations with a single click. The Template I2V mode provides an additional layer of consistency: when generating video variants for A/B testing or batch processing, the fixed motion templates ensure that differences in output are driven by the input image rather than random motion seeds, making comparison and iteration more reliable. + +The videos generated by PixVerse can be further processed in ComfyUI by chaining with other nodes. Common post-processing workflows include applying video filters via VHS nodes, compositing PixVerse output over a background video, extracting frames for frame-by-frame editing, or using audio-reactive nodes to sync generated video with a soundtrack. This composability makes PixVerse a versatile building block rather than an isolated tool — it fits naturally into larger production pipelines spanning concept development, asset creation, and final rendering. + + + + +## Key capabilities + +- **Text-to-Video** — Generate original videos from text prompts, ideal for concept visualization, storyboarding, and rapid creative exploration. +- **Image-to-Video** — Animate a reference image while preserving subject identity and scene composition, bringing static visuals to life with natural motion. +- **Template Image-to-Video** — Apply pre-defined motion templates to input images for predictable, repeatable animation styles suitable for production pipelines. + +## Available workflows + +### Text-to-Video + +Create original videos from text prompts. Describe the scene, characters, and action — PixVerse generates a corresponding video clip with cinematic quality. + + + + Run the text-to-video workflow instantly on Comfy Cloud. + + + Download the workflow JSON file for local use. + + + +### Image-to-Video + +Take a reference image and generate a video that animates it while preserving the original subject, style, and composition. + + + + Run the image-to-video workflow instantly on Comfy Cloud. + + + Download the workflow JSON file for local use. + + + +### Template Image-to-Video + +Apply a pre-defined motion template to your input image for predictable, repeatable animation results. This mode reduces variability and is ideal for consistent batch processing. + + + + Run the template image-to-video workflow instantly on Comfy Cloud. + + + Download the workflow JSON file for local use. + + diff --git a/tutorials/partner-nodes/quiver/quiver.mdx b/tutorials/partner-nodes/quiver/quiver.mdx new file mode 100644 index 000000000..8817d2180 --- /dev/null +++ b/tutorials/partner-nodes/quiver/quiver.mdx @@ -0,0 +1,59 @@ +--- +title: "Quiver AI API Node SVG Generation ComfyUI Official Examples" +description: "This guide covers how to use the Quiver partner nodes in ComfyUI for AI-powered SVG vector graphics generation from text prompts and reference images." +sidebarTitle: "Quiver AI" +--- + +import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; +import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; + +Quiver is an AI-powered vector graphics platform that specializes in generating high-quality SVG (Scalable Vector Graphics) from text descriptions and reference images. Now integrated as partner nodes in ComfyUI, Quiver enables you to create clean, scalable vector artwork directly within your workflows — no manual vector editing required. + +The Quiver node suite offers two generation modes. **Text-to-SVG** lets you describe the vector graphic you want in natural language, and Quiver generates a fully editable SVG file. This is ideal for creating icons, illustrations, logos, diagrams, and UI elements without needing graphic design skills. Just describe what you want — "a minimalist mountain landscape silhouette," "a three-icon set for settings, profile, and notifications" — and the SVG is generated ready for use in web design, print, or further editing in tools like Figma or Illustrator. **Image-to-SVG** takes an existing raster image (PNG, JPG, etc.) as input and converts it into a clean vector representation. This process traces the shapes, colors, and gradients from the source image and reconstructs them as scalable vector paths, making it perfect for converting hand-drawn sketches, bitmap logos, or photographed artwork into resolution-independent SVGs. + +Quiver supports three model variants that offer different trade-offs between generation quality and speed. **Arrow 1.1** is the standard model — fast and efficient, suitable for most use cases. **Arrow 1.1 Max** is the high-quality variant that produces more detailed and accurate vector output, ideal for complex illustrations where visual fidelity matters most. **Arrow Preview** offers early access to experimental features and upcoming model improvements. Each variant can be selected directly from the node's model parameter, giving you the flexibility to choose the right balance of speed and quality for your specific task. + +Vector graphics generated by Quiver have significant advantages over raster images in many design contexts. SVGs are resolution-independent — they look sharp at any display size, from mobile icons to billboards. They also have smaller file sizes for geometric artwork, can be styled with CSS, animated with SMIL or JavaScript, and embedded directly into web pages without additional HTTP requests. This makes Quiver's output immediately useful for web development, UI/UX design, screen printing, laser engraving, and any application where scalable resolution and small file footprints matter. + +In practice, Quiver integrates naturally with typical design workflows in ComfyUI. For Text-to-SVG, try prompting with specific style cues — "flat design icon," "line art illustration," "filled silhouette" — to guide the model toward the desired vector aesthetic. For Image-to-SVG, best results come from source images with clear shapes, high contrast, and limited color palettes; complex photographic images with many gradients may produce more path segments and larger file sizes. The Arrow 1.1 Max model is recommended when the SVG will be scaled up significantly (e.g., for large-format printing), while Arrow 1.1 is sufficient for web assets and icons where speed is a priority. + +Once generated, SVGs can be output through ComfyUI's standard save nodes or piped into downstream processing. You can combine Quiver-generated SVGs with raster nodes — for example, rasterizing an SVG at a specific resolution and feeding it into a GPT-Image-1 or Magnific enhancement node — creating hybrid workflows that leverage the strengths of both vector precision and AI image generation. + + + + +## Key capabilities + +- **Text-to-SVG** — Generate fully editable SVG vector graphics from natural language descriptions; perfect for icons, logos, illustrations, and UI elements. +- **Image-to-SVG** — Convert raster images (PNG, JPG) into clean, scalable SVG vector graphics by tracing shapes, colors, and gradients. +- **Arrow 1.1** — Standard model offering fast and efficient SVG generation suitable for most use cases. +- **Arrow 1.1 Max** — High-quality variant for detailed and accurate vector output on complex illustrations. +- **Arrow Preview** — Early access to experimental features and upcoming model improvements. + +## Available workflows + +### Text-to-SVG + +Describe the vector graphic you want in natural language, and Quiver generates a clean, editable SVG. Use it for creating icons, logos, diagrams, illustrations, and UI components. + + + + Run the text-to-SVG workflow instantly on Comfy Cloud. + + + Download the workflow JSON file for local use. + + + +### Image-to-SVG + +Convert an existing raster image into a clean, scalable SVG vector graphic. Shapes, colors, and gradients from the source are traced and reconstructed as vector paths. + + + + Run the image-to-SVG workflow instantly on Comfy Cloud. + + + Download the workflow JSON file for local use. + + diff --git a/tutorials/partner-nodes/topaz/topaz.mdx b/tutorials/partner-nodes/topaz/topaz.mdx new file mode 100644 index 000000000..f923550d3 --- /dev/null +++ b/tutorials/partner-nodes/topaz/topaz.mdx @@ -0,0 +1,69 @@ +--- +title: "Image & Video Enhancement with Topaz Labs" +description: "Upscale, enhance, and restore images and video using Topaz Labs AI models — including Reimagine creative upscaling, face enhancement, Starlight diffusion upscale, and Apollo frame interpolation — all within ComfyUI." +sidebarTitle: "Topaz Labs" +--- + +import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; +import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; + +Topaz Labs has long been the gold standard in AI-powered image and video enhancement, trusted by professional photographers, videographers, and post-production studios worldwide. Now integrated into ComfyUI, Topaz's suite of models brings enterprise-grade upscaling, restoration, and enhancement directly into your visual workflow — no more round-tripping between applications. + +The integration covers two primary domains: **image enhancement** and **video enhancement**. On the image side, **api_topaz_image_enhance** provides general-purpose AI upscaling and denoising, while the **Starlight** diffusion-based upscaler (utility_topaz_landscape_upscaler) specializes in landscape photography — it uses a diffusion process to hallucinate realistic detail beyond simple pixel interpolation, producing stunning 8K-level output from lower-resolution sources. The **Reimagine** mode takes this further by allowing creative reinterpretation: you can guide the upscaling with a text prompt, transforming a simple photo into a completely restyled or enhanced version. For illustration and anime art, **utility_topaz_illustration_upscale** applies models optimized for flat colors, line art, and non-photographic content, preserving crisp edges and smooth gradients. + +On the video side, **api_topaz_video_enhance** brings Topaz's video AI to ComfyUI, enabling frame-by-frame intelligent upscaling, denoising, deinterlacing, and stabilization. The **Apollo** frame interpolation model (included in the video pipeline) generates intermediate frames to create smooth slow-motion or increase frame rates, making it ideal for converting 24fps footage to 60fps or beyond. + +One standout feature is **face enhancement** — Topaz's specialized facial detail reconstruction model that detects and enhances faces within both images and video frames. This is particularly valuable for portrait photography, interview footage, and any content where human faces are the focus, as traditional upscaling often leaves facial features blurry or unnatural. + +All Topaz nodes in ComfyUI are powered by the Topaz API via ComfyCloud, meaning no local GPU is required for computation. You can chain Topaz enhancement with other ComfyUI nodes for a complete pipeline — generate with AI, enhance with Topaz, and composite in a single session. + +## Key capabilities + +- **Reimagine creative upscaling** — Guide the upscaling process with text prompts to creatively reinterpret and enhance images, not just enlarge them. +- **Face enhancement** — Dedicated facial detail reconstruction model that detects and sharpens faces in images and video, preserving natural skin texture and expression. +- **8K output support** — Upscale images and video frames to ultra-high 8K resolution while maintaining sharpness and visual coherence. +- **Starlight diffusion upscaler** — Diffusion-based upscaling optimized for landscape photography, hallucinating realistic detail like foliage, rock textures, and clouds. +- **Apollo frame interpolation** — Generate smooth intermediate frames between existing video frames for slow-motion effects or frame rate conversion. +- **Illustration-optimized upscale** — Specialized model for flat-colored artwork, line art, and digital illustrations that preserves hard edges and clean gradients. +- **Video enhancement pipeline** — Comprehensive video processing including denoising, deinterlacing, stabilization, and intelligent frame-by-frame upscaling. + + + + +## Available workflows + +### Image Enhancement + +General-purpose AI image upscaling and denoising using Topaz's core enhancement model. Suitable for photographs, renders, and most image types. + + + Launch the image enhancement workflow on ComfyCloud + Download the workflow JSON for local ComfyUI + + +### Video Enhancement + +Frame-by-frame AI video upscaling, denoising, and stabilization. Supports deinterlacing and Apollo frame interpolation for smooth slow-motion. + + + Launch the video enhancement workflow on ComfyCloud + Download the workflow JSON for local ComfyUI + + +### Landscape Upscaler (Starlight) + +Diffusion-based upscaling specialized for landscape photography. Uses the Starlight model to hallucinate realistic natural detail at up to 8K resolution. + + + Launch the Starlight landscape upscaler on ComfyCloud + Download the workflow JSON for local ComfyUI + + +### Illustration Upscale + +Optimized upscaling for digital illustrations, anime art, and flat-color artwork. Preserves hard edges, crisp lines, and smooth color gradients. + + + Launch the illustration upscale workflow on ComfyCloud + Download the workflow JSON for local ComfyUI + diff --git a/tutorials/partner-nodes/veo/veo.mdx b/tutorials/partner-nodes/veo/veo.mdx new file mode 100644 index 000000000..faee76908 --- /dev/null +++ b/tutorials/partner-nodes/veo/veo.mdx @@ -0,0 +1,57 @@ +--- +title: "Video Generation with Google Veo" +description: "Generate cinematic, high-resolution videos using Google DeepMind's Veo models — Veo 3.1 and Veo 2.0 — directly within ComfyUI." +sidebarTitle: "Veo" +--- + +import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; +import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; + +Google DeepMind's Veo family represents the state of the art in AI video generation, now deeply integrated into ComfyUI through a suite of dedicated workflow nodes. The flagship model, **Veo 3.1**, delivers cinema-grade 4K video with native audio — a breakthrough that brings full-production-quality video generation into a single pipeline. Whether you're creating short films, product demos, social media content, or concept visualizations, Veo 3.1 produces realistic motion, consistent character identities, and scene-coherent narratives that rival traditional film production. + +Alongside Veo 3.1, **Veo 2.0** offers a lighter, faster text-to-video and image-to-video alternative that balances quality with speed, making it ideal for rapid prototyping and iterative creative workflows. Veo 2.0's image-to-video capability (api_veo2_i2v) lets you animate still images with natural motion, breathing life into concept art, photography, or AI-generated visuals. + +All Veo nodes are powered by Google's infrastructure and accessed via the ComfyCloud platform. The tight integration means you can chain Veo generation with ComfyUI's post-processing nodes — upscaling, color grading, audio mixing, and compositing — all in a single visual workflow. This eliminates the need to export and re-import media between different tools, dramatically accelerating the video production pipeline. + +Key differentiators of Veo in ComfyUI include: native audio generation synchronized with video, high-fidelity 4K output that preserves fine details and textures, temporal consistency across frames that eliminates flickering and jarring transitions, and support for detailed prompt-guided cinematography including camera angles, lighting conditions, and scene composition. For creators who need production-ready video from text or image prompts, Veo represents the most advanced AI video generation option available today. + +## Key capabilities + +- **Cinema-grade 4K video** — Veo 3.1 generates 4K resolution video with crisp detail, natural lighting, and realistic physics, suitable for professional film and advertising use. +- **Native audio sync** — Video output includes synchronized audio, eliminating the need for separate audio generation and manual alignment in post-production. +- **Text-to-video (Veo 3.1 & 2.0)** — Generate high-quality video directly from text prompts, with support for detailed scene direction, camera movement, and visual style control. +- **Image-to-video (Veo 2.0)** — Animate static images into short video clips, preserving the original composition while adding natural motion. +- **Optimized speed tier (api_veo3_fast)** — A faster inference variant of Veo 3.1 that trades slight quality for significantly reduced generation time, ideal for iteration and draft production. +- **Seamless ComfyUI pipeline** — Full integration allows chaining Veo output with ComfyUI's existing upscaling, audio, and compositing nodes without file export steps. + + + + +## Available workflows + +### Veo 3.1 — Text-to-Video (Standard) + +Generate full-quality 4K video with native audio from a text prompt. Best for final production output where quality matters most. + + + Launch the full-quality Veo 3.1 workflow on ComfyCloud + Download the workflow JSON for local ComfyUI + + +### Veo 3.1 — Text-to-Video (Fast) + +A lightweight variant of Veo 3.1 optimized for speed. Use this for rapid ideation, draft generation, and iterative prompt refinement. + + + Launch the fast Veo 3.1 workflow on ComfyCloud + Download the workflow JSON for local ComfyUI + + +### Veo 2.0 — Image-to-Video + +Animate a static image into a short video clip. Great for breathing life into concept art, AI-generated images, or photographs. + + + Launch the Veo 2.0 I2V workflow on ComfyCloud + Download the workflow JSON for local ComfyUI + diff --git a/tutorials/partner-nodes/vidu/vidu.mdx b/tutorials/partner-nodes/vidu/vidu.mdx new file mode 100644 index 000000000..d8f58f3b9 --- /dev/null +++ b/tutorials/partner-nodes/vidu/vidu.mdx @@ -0,0 +1,135 @@ +--- +title: "Vidu - AI video generation" +description: "Generate high-quality video from text, images, and reference materials using Vidu's Q1, Q2, and Q3 model generations in ComfyUI" +sidebarTitle: "Vidu" +--- + +import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; +import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; + +Vidu is an AI video generation platform developed by Shengshu Technology, known for its fast inference speed and high visual quality across multiple model generations. Now available inside ComfyUI, Vidu offers three model families — Q1, Q2, and Q3 — each tailored to different creative needs. Q1 provides the core foundation for text-to-video, image-to-video, reference-to-video, start-end-to-video, and video extension workflows. Q2 brings 3x faster generation, support for up to 7 reference images, stronger subject consistency, and cinematic camera language controls including pan, zoom, and rotation. Q3 is the latest generation, delivering the highest visual fidelity with exceptional performance in anime and 2D animation styles, and can generate a 4-second video clip in as fast as 10 seconds. Whether you need quick drafts, polished character-driven narratives, or stylized animation sequences, Vidu's tiered model lineup gives you the flexibility to balance speed, quality, and creative control. + +## Key capabilities + +- **Multi-model generation** — Choose between Q1 (foundation), Q2 (fast + multi-reference + camera control), and Q3 (latest narrative + anime/2D focus) to match your workflow +- **Fast inference** — Q3 generates 4-second clips in as fast as 10 seconds; Q2 delivers 3x the speed of Q1 +- **Rich input modes** — Text-to-video, image-to-video (with optional prompt), reference-to-video (up to 7 images), start-end-to-video, and video extension +- **Camera language (Q2)** — Pan, zoom, rotation, and other cinematic moves for precise shot direction +- **Exceptional anime & 2D animation (Q3)** — Industry-leading output quality for animation production pipelines +- **Subject consistency** — Maintain character and scene identity across multi-reference inputs + + + + +## Available workflows + +### Q1 — Foundation generation + +The first-generation Vidu model handles all core generation types: text-to-video, image-to-video, reference-to-video, start-end-to-video, and video extension. Reliable for general-purpose production where speed and expressiveness are balanced. + + + + Try the Q1 Text-to-Video workflow instantly on Comfy Cloud. + + + Download the workflow JSON. + + + + + + Try the Q1 Image-to-Video workflow instantly on Comfy Cloud. + + + Download the workflow JSON. + + + + + + Try the Q1 Reference-to-Video workflow instantly on Comfy Cloud. + + + Download the workflow JSON. + + + + + + Try the Q1 Start-End-to-Video workflow instantly on Comfy Cloud. + + + Download the workflow JSON. + + + + + + Try the Q1 Video Extension workflow instantly on Comfy Cloud. + + + Download the workflow JSON. + + + +### Q2 — Fast generation with camera control and multi-reference + +Q2 builds on Q1 with 3x faster inference, support for up to 7 reference images, enhanced subject consistency, and cinematic camera language controls including pan, zoom, and rotation. Ideal for iterative workflows and shot-level direction. Available as text-to-video, image-to-video, reference-to-video, and first-last-frame-to-video. + + + + Try the Q2 Text-to-Video workflow instantly on Comfy Cloud. + + + Download the workflow JSON. + + + + + + Try the Q2 Image-to-Video workflow instantly on Comfy Cloud. + + + Download the workflow JSON. + + + + + + Try the Q2 Reference-to-Video workflow instantly on Comfy Cloud. + + + Download the workflow JSON. + + + + + + Try the Q2 First-Last-Frame-to-Video workflow instantly on Comfy Cloud. + + + Download the workflow JSON. + + + +### Q3 — Latest generation for narrative and animation + +Vidu Q3 is the newest and most capable model generation. It delivers the highest visual quality with a particular strength in anime and 2D animation styles, making it the top choice for animation production. Q3 also introduces improved narrative understanding for more coherent storytelling. Generation is remarkably fast — a 4-second clip can be produced in as little as 10 seconds. Available as text-to-video and image-to-video. + + + + Try the Q3 Text-to-Video workflow instantly on Comfy Cloud. + + + Download the workflow JSON. + + + + + + Try the Q3 Image-to-Video workflow instantly on Comfy Cloud. + + + Download the workflow JSON. + + diff --git a/tutorials/partner-nodes/wavespeed/wavespeed.mdx b/tutorials/partner-nodes/wavespeed/wavespeed.mdx new file mode 100644 index 000000000..9611c2aad --- /dev/null +++ b/tutorials/partner-nodes/wavespeed/wavespeed.mdx @@ -0,0 +1,59 @@ +--- +title: "Image & Video Enhancement with WaveSpeed" +description: "Upscale video and images, and restore old photos using WaveSpeed AI models — FLSH-VSR video upscaling, SeedVR2 image restoration, and general image upscaling, all within ComfyUI." +sidebarTitle: "WaveSpeed" +--- + +import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; +import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; + +WaveSpeed is an emerging AI media enhancement platform that specializes in high-performance upscaling and restoration models, now available as native workflow nodes in ComfyUI. The integration brings three distinct capabilities to your toolbox: **FLSH-VSR video upscaling**, **general image upscaling**, and the **SeedVR2 AI image fix** for restoring damaged or low-quality photographs. All processing is handled via the ComfyCloud API, requiring no local GPU resources. + +**FLSH-VSR** (api_wavespeed_flshvsr_video_upscale) is WaveSpeed's flagship video super-resolution model. It takes low-resolution video input and intelligently reconstructs high-resolution output frame by frame, with particular strength in temporal coherence — meaning the enhanced frames maintain consistent detail and lighting across cuts and movements. Unlike simpler frame-independent upscalers that can introduce flickering or jitter, FLSH-VSR analyzes motion vectors between frames to produce smooth, realistic results. This makes it ideal for upscaling archived video footage, improving AI-generated video quality, or converting SD content to HD or beyond. + +For standalone image upscaling, **api_wavespped_image_upscale** (note the typo in the template identifier, preserved for consistency) provides a robust general-purpose model. It handles common upscaling scenarios — enlarging AI-generated images, improving photograph resolution, and cleaning up compression artifacts — with natural-looking detail reconstruction. The model works across diverse content types including photography, digital art, renders, and screenshots. + +The **SeedVR2 AI image fix** (api_wavespped_seedvr2_ai_image_fix) is a specialized restoration model designed for recovering damaged, degraded, or extremely low-quality images. It goes beyond simple upscaling by performing intelligent reconstruction of missing or corrupted regions: it can remove noise patterns, repair compression blocking, fix color degradation, and reconstruct lost facial details in old photographs. This makes it an excellent choice for photo restoration projects, recovering detail from heavily compressed web images, or salvaging generations that turned out blurry or artifact-ridden. SeedVR2 applies a diffusion-based approach to "imagine" the most likely clean version of the input, producing results that look natural rather than artificially sharpened. + +All three WaveSpeed nodes are designed as straightforward single-pass operations that integrate easily into larger ComfyUI pipelines. Whether you're enhancing final output or preprocessing source material for further generation, WaveSpeed provides reliable, high-quality results with minimal configuration. + +## Key capabilities + +- **FLSH-VSR video upscaling** — Temporal-coherent AI video super-resolution that maintains consistent detail across frames, ideal for archiving, AI video enhancement, and SD-to-HD conversion. +- **General image upscaling** — Multi-purpose image enlargement and quality improvement covering photographs, digital art, renders, and screenshots. +- **SeedVR2 AI image fix** — Diffusion-based restoration model for recovering damaged, degraded, or low-quality images, including old photo repair, noise and artifact removal, and lost detail reconstruction. +- **No local GPU required** — All processing runs on ComfyCloud's infrastructure, keeping your local hardware free for generation and composition. +- **Easy pipeline integration** — Simple drop-in nodes that work as a final enhancement step or intermediate preprocessing stage in any ComfyUI workflow. +- **Broad content compatibility** — Models handle diverse input types including photographs, AI art, video frames, and compressed web images. + + + + +## Available workflows + +### FLSH-VSR Video Upscaling + +Upscale low-resolution video to higher quality using WaveSpeed's temporal-aware super-resolution model. Maintains frame-to-frame consistency for smooth, realistic output. + + + Launch the FLSH-VSR video upscale workflow on ComfyCloud + Download the workflow JSON for local ComfyUI + + +### Image Upscaling + +General-purpose AI image upscaling for photographs, renders, and digital artwork. Enlarges and enhances while preserving natural detail. + + + Launch the image upscale workflow on ComfyCloud + Download the workflow JSON for local ComfyUI + + +### SeedVR2 AI Image Fix + +Restore damaged, degraded, or low-quality images using WaveSpeed's diffusion-based restoration model. Repairs noise, artifacts, color degradation, and missing facial details. + + + Launch the SeedVR2 image fix workflow on ComfyCloud + Download the workflow JSON for local ComfyUI + From 43067bea31473749c6675d71a9c75b19fa9c06c8 Mon Sep 17 00:00:00 2001 From: linmoumou Date: Mon, 27 Apr 2026 00:45:25 +0800 Subject: [PATCH 2/9] docs: register 11 new partner-node pages in docs.json (en, zh, ja) --- docs.json | 400 ++++++++++++++++++++++++++++++++++++++++-------------- 1 file changed, 299 insertions(+), 101 deletions(-) diff --git a/docs.json b/docs.json index 6518dcbb7..ad2b30e3b 100644 --- a/docs.json +++ b/docs.json @@ -318,6 +318,12 @@ "tutorials/partner-nodes/bytedance/seedream-5-lite" ] }, + { + "group": "ElevenLabs", + "pages": [ + "partner-nodes/elevenlabs/elevenlabs" + ] + }, { "group": "Google", "pages": [ @@ -326,6 +332,42 @@ "tutorials/partner-nodes/google/nano-banana-2" ] }, + { + "group": "Grok", + "pages": [ + "partner-nodes/grok/grok" + ] + }, + { + "group": "HitPaw", + "pages": [ + "partner-nodes/hitpaw/hitpaw" + ] + }, + { + "group": "Magnific", + "pages": [ + "partner-nodes/magnific/magnific" + ] + }, + { + "group": "MiniMax", + "pages": [ + "partner-nodes/minimax/minimax" + ] + }, + { + "group": "PixVerse", + "pages": [ + "partner-nodes/pixverse/pixverse" + ] + }, + { + "group": "Quiver", + "pages": [ + "partner-nodes/quiver/quiver" + ] + }, { "group": "Stability AI", "pages": [ @@ -391,6 +433,12 @@ "tutorials/partner-nodes/rodin/model-generation" ] }, + { + "group": "Topaz", + "pages": [ + "partner-nodes/topaz/topaz" + ] + }, { "group": "Tripo", "pages": [ @@ -421,6 +469,18 @@ "tutorials/partner-nodes/reve/reve-image" ] }, + { + "group": "Veo", + "pages": [ + "partner-nodes/veo/veo" + ] + }, + { + "group": "Vidu", + "pages": [ + "partner-nodes/vidu/vidu" + ] + }, { "group": "Wan", "pages": [ @@ -432,6 +492,12 @@ "pages": [ "tutorials/partner-nodes/sonilo/video-to-music" ] + }, + { + "group": "WaveSpeed", + "pages": [ + "partner-nodes/wavespeed/wavespeed" + ] } ] } @@ -2124,19 +2190,19 @@ "language": "zh", "tabs": [ { - "tab": "开始使用", + "tab": "\u5f00\u59cb\u4f7f\u7528", "pages": [ { - "group": "开始使用", + "group": "\u5f00\u59cb\u4f7f\u7528", "pages": [ "zh/index", { - "group": "本地安装", + "group": "\u672c\u5730\u5b89\u88c5", "icon": "download", "pages": [ "zh/installation/system_requirements", { - "group": "Desktop(推荐)", + "group": "Desktop(\u63a8\u8350)", "pages": [ "zh/installation/desktop/windows", "zh/installation/desktop/macos", @@ -2150,7 +2216,7 @@ }, "zh/get_started/cloud", { - "group": "安装自定义节点", + "group": "\u5b89\u88c5\u81ea\u5b9a\u4e49\u8282\u70b9", "icon": "puzzle-piece", "pages": [ "zh/installation/install_custom_node", @@ -2160,7 +2226,7 @@ "zh/manager/overview", "zh/manager/install", { - "group": "自定义节点管理", + "group": "\u81ea\u5b9a\u4e49\u8282\u70b9\u7ba1\u7406", "pages": [ "zh/manager/pack-management", "zh/manager/legacy-ui" @@ -2176,7 +2242,7 @@ ] }, { - "group": "基础概念", + "group": "\u57fa\u7840\u6982\u5ff5", "pages": [ "zh/development/core-concepts/workflow", "zh/development/core-concepts/nodes", @@ -2188,7 +2254,7 @@ ] }, { - "group": "界面指南", + "group": "\u754c\u9762\u6307\u5357", "pages": [ "zh/interface/overview", "zh/interface/app-mode", @@ -2199,7 +2265,7 @@ "zh/interface/features/partial-execution", "zh/interface/features/node-docs", { - "group": "ComfyUI 设置", + "group": "ComfyUI \u8bbe\u7f6e", "icon": "gear", "pages": [ "zh/interface/settings/overview", @@ -2218,7 +2284,7 @@ ] }, { - "group": "Cloud 专属功能", + "group": "Cloud \u4e13\u5c5e\u529f\u80fd", "icon": "cloud", "pages": [ "zh/cloud/share-workflow", @@ -2228,11 +2294,11 @@ ] }, { - "group": "教程示例", + "group": "\u6559\u7a0b\u793a\u4f8b", "icon": "book", "pages": [ { - "group": "基础示例", + "group": "\u57fa\u7840\u793a\u4f8b", "pages": [ "zh/tutorials/basic/text-to-image", "zh/tutorials/basic/image-to-image", @@ -2316,7 +2382,7 @@ ] }, { - "group": "视频", + "group": "\u89c6\u9891", "pages": [ { "group": "LTX", @@ -2327,7 +2393,7 @@ ] }, { - "group": "万相视频", + "group": "\u4e07\u76f8\u89c6\u9891", "pages": [ "zh/tutorials/video/wan/wan2_2", "zh/tutorials/video/wan/wan2-2-animate", @@ -2352,7 +2418,7 @@ ] }, { - "group": "腾讯混元", + "group": "\u817e\u8baf\u6df7\u5143", "pages": [ "zh/tutorials/video/hunyuan/hunyuan-video", "zh/tutorials/video/hunyuan/hunyuan-video-1-5" @@ -2373,7 +2439,7 @@ ] }, { - "group": "音频", + "group": "\u97f3\u9891", "pages": [ { "group": "ACE-Step", @@ -2394,7 +2460,7 @@ ] }, { - "group": "合作伙伴节点", + "group": "\u5408\u4f5c\u4f19\u4f34\u8282\u70b9", "pages": [ "zh/tutorials/partner-nodes/overview", "zh/tutorials/partner-nodes/faq", @@ -2414,6 +2480,12 @@ "zh/tutorials/partner-nodes/bytedance/seedream-5-lite" ] }, + { + "group": "ElevenLabs", + "pages": [ + "zh/tutorials/partner-nodes/elevenlabs/elevenlabs" + ] + }, { "group": "Google", "pages": [ @@ -2422,6 +2494,42 @@ "zh/tutorials/partner-nodes/google/nano-banana-2" ] }, + { + "group": "Grok", + "pages": [ + "zh/tutorials/partner-nodes/grok/grok" + ] + }, + { + "group": "HitPaw", + "pages": [ + "zh/tutorials/partner-nodes/hitpaw/hitpaw" + ] + }, + { + "group": "Magnific", + "pages": [ + "zh/tutorials/partner-nodes/magnific/magnific" + ] + }, + { + "group": "MiniMax", + "pages": [ + "zh/tutorials/partner-nodes/minimax/minimax" + ] + }, + { + "group": "PixVerse", + "pages": [ + "zh/tutorials/partner-nodes/pixverse/pixverse" + ] + }, + { + "group": "Quiver", + "pages": [ + "zh/tutorials/partner-nodes/quiver/quiver" + ] + }, { "group": "Stability AI", "pages": [ @@ -2487,6 +2595,12 @@ "zh/tutorials/partner-nodes/rodin/model-generation" ] }, + { + "group": "Topaz", + "pages": [ + "zh/tutorials/partner-nodes/topaz/topaz" + ] + }, { "group": "Tripo", "pages": [ @@ -2517,6 +2631,18 @@ "zh/tutorials/partner-nodes/reve/reve-image" ] }, + { + "group": "Veo", + "pages": [ + "zh/tutorials/partner-nodes/veo/veo" + ] + }, + { + "group": "Vidu", + "pages": [ + "zh/tutorials/partner-nodes/vidu/vidu" + ] + }, { "group": "Wan", "pages": [ @@ -2528,6 +2654,12 @@ "pages": [ "zh/tutorials/partner-nodes/sonilo/video-to-music" ] + }, + { + "group": "WaveSpeed", + "pages": [ + "zh/tutorials/partner-nodes/wavespeed/wavespeed" + ] } ] } @@ -2537,11 +2669,11 @@ ] }, { - "tab": "内置节点", + "tab": "\u5185\u7f6e\u8282\u70b9", "pages": [ "zh/built-in-nodes/overview", { - "group": "节点", + "group": "\u8282\u70b9", "pages": [ { "group": "3D", @@ -3024,7 +3156,7 @@ ] }, { - "group": "图像", + "group": "\u56fe\u50cf", "pages": [ { "group": "BFL", @@ -3177,7 +3309,7 @@ ] }, { - "group": "加载器", + "group": "\u52a0\u8f7d\u5668", "pages": [ { "group": "Video Models", @@ -3207,7 +3339,7 @@ ] }, { - "group": "图像", + "group": "\u56fe\u50cf", "pages": [ { "group": "Animation", @@ -3348,7 +3480,7 @@ ] }, { - "group": "实用工具", + "group": "\u5b9e\u7528\u5de5\u5177", "pages": [ { "group": "Primitive", @@ -3390,7 +3522,7 @@ ] }, { - "group": "条件", + "group": "\u6761\u4ef6", "pages": [ { "group": "3D Models", @@ -3524,7 +3656,7 @@ ] }, { - "group": "模型补丁", + "group": "\u6a21\u578b\u8865\u4e01", "pages": [ { "group": "Chroma Radiance", @@ -3549,7 +3681,7 @@ ] }, { - "group": "潜变量", + "group": "\u6f5c\u53d8\u91cf", "pages": [ { "group": "3D", @@ -3680,7 +3812,7 @@ ] }, { - "group": "采样", + "group": "\u91c7\u6837", "pages": [ { "group": "Custom Sampling", @@ -3773,7 +3905,7 @@ ] }, { - "group": "高级", + "group": "\u9ad8\u7ea7", "pages": [ { "group": "Attention Experiments", @@ -4038,7 +4170,7 @@ ] }, { - "tab": "开发", + "tab": "\u5f00\u53d1", "pages": [ "zh/development/overview", { @@ -4071,12 +4203,12 @@ ] }, { - "group": "开发自定义节点", + "group": "\u5f00\u53d1\u81ea\u5b9a\u4e49\u8282\u70b9", "pages": [ "zh/custom-nodes/overview", "zh/custom-nodes/walkthrough", { - "group": "后端", + "group": "\u540e\u7aef", "icon": "python", "pages": [ "zh/custom-nodes/backend/server_overview", @@ -4121,7 +4253,7 @@ ] }, { - "group": "注册表(Registry)", + "group": "\u6ce8\u518c\u8868(Registry)", "pages": [ "zh/registry/overview", "zh/registry/publishing", @@ -4132,17 +4264,17 @@ ] }, { - "group": "规范", + "group": "\u89c4\u8303", "pages": [ { - "group": "工作流 JSON", + "group": "\u5de5\u4f5c\u6d41 JSON", "pages": [ "zh/specs/workflow_json", "zh/specs/workflow_json_0.4" ] }, { - "group": "节点定义", + "group": "\u8282\u70b9\u5b9a\u4e49", "pages": [ "zh/specs/nodedef_json", "zh/specs/nodedef_json_1_0" @@ -4153,11 +4285,11 @@ ] }, { - "tab": "支持", + "tab": "\u652f\u6301", "pages": [ "zh/support/contact-support", { - "group": "账户管理", + "group": "\u8d26\u6237\u7ba1\u7406", "icon": "user", "pages": [ "zh/account/create-account", @@ -4166,10 +4298,10 @@ ] }, { - "group": "账单支持", + "group": "\u8d26\u5355\u652f\u6301", "pages": [ { - "group": "订阅", + "group": "\u8ba2\u9605", "pages": [ "zh/support/subscription/subscribing", "zh/support/subscription/managing", @@ -4178,7 +4310,7 @@ ] }, { - "group": "支付", + "group": "\u652f\u4ed8", "pages": [ "zh/support/payment/accepted-payment-methods", "zh/support/payment/editing-payment-information", @@ -4191,7 +4323,7 @@ ] }, { - "group": "故障排除", + "group": "\u6545\u969c\u6392\u9664", "icon": "bug", "pages": [ "zh/troubleshooting/overview", @@ -4200,7 +4332,7 @@ ] }, { - "group": "社区", + "group": "\u793e\u533a", "pages": [ "zh/community/contributing", "zh/community/links" @@ -4213,7 +4345,7 @@ "openapi": "https://api.comfy.org/openapi" }, { - "tab": "Cloud API 参考文档", + "tab": "Cloud API \u53c2\u8003\u6587\u6863", "openapi": { "source": "openapi-cloud.yaml", "directory": "zh/api-reference/cloud" @@ -4225,19 +4357,19 @@ "language": "ja", "tabs": [ { - "tab": "はじめに", + "tab": "\u306f\u3058\u3081\u306b", "pages": [ { - "group": "はじめに", + "group": "\u306f\u3058\u3081\u306b", "pages": [ "ja/index", { - "group": "ローカルインストール", + "group": "\u30ed\u30fc\u30ab\u30eb\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb", "icon": "download", "pages": [ "ja/installation/system_requirements", { - "group": "デスクトップ（推奨）", + "group": "\u30c7\u30b9\u30af\u30c8\u30c3\u30d7\uff08\u63a8\u5968\uff09", "pages": [ "ja/installation/desktop/windows", "ja/installation/desktop/macos", @@ -4251,7 +4383,7 @@ }, "ja/get_started/cloud", { - "group": "カスタムノードのインストール", + "group": "\u30ab\u30b9\u30bf\u30e0\u30ce\u30fc\u30c9\u306e\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb", "icon": "puzzle-piece", "pages": [ "ja/installation/install_custom_node", @@ -4261,7 +4393,7 @@ "ja/manager/overview", "ja/manager/install", { - "group": "カスタムノード管理", + "group": "\u30ab\u30b9\u30bf\u30e0\u30ce\u30fc\u30c9\u7ba1\u7406", "pages": [ "ja/manager/pack-management", "ja/manager/legacy-ui" @@ -4277,7 +4409,7 @@ ] }, { - "group": "基本概念", + "group": "\u57fa\u672c\u6982\u5ff5", "pages": [ "ja/development/core-concepts/workflow", "ja/development/core-concepts/nodes", @@ -4289,7 +4421,7 @@ ] }, { - "group": "インターフェースガイド", + "group": "\u30a4\u30f3\u30bf\u30fc\u30d5\u30a7\u30fc\u30b9\u30ac\u30a4\u30c9", "pages": [ "ja/interface/overview", "ja/interface/app-mode", @@ -4300,7 +4432,7 @@ "ja/interface/features/partial-execution", "ja/interface/features/node-docs", { - "group": "ComfyUI 設定", + "group": "ComfyUI \u8a2d\u5b9a", "icon": "gear", "pages": [ "ja/interface/settings/overview", @@ -4319,7 +4451,7 @@ ] }, { - "group": "Cloud 専用機能", + "group": "Cloud \u5c02\u7528\u6a5f\u80fd", "icon": "cloud", "pages": [ "ja/cloud/share-workflow", @@ -4329,11 +4461,11 @@ ] }, { - "group": "チュートリアル", + "group": "\u30c1\u30e5\u30fc\u30c8\u30ea\u30a2\u30eb", "icon": "book", "pages": [ { - "group": "基本チュートリアル", + "group": "\u57fa\u672c\u30c1\u30e5\u30fc\u30c8\u30ea\u30a2\u30eb", "pages": [ "ja/tutorials/basic/text-to-image", "ja/tutorials/basic/image-to-image", @@ -4417,7 +4549,7 @@ ] }, { - "group": "動画", + "group": "\u52d5\u753b", "pages": [ { "group": "LTX", @@ -4474,7 +4606,7 @@ ] }, { - "group": "オーディオ", + "group": "\u30aa\u30fc\u30c7\u30a3\u30aa", "pages": [ { "group": "ACE-Step", @@ -4495,7 +4627,7 @@ ] }, { - "group": "パートナーノード", + "group": "\u30d1\u30fc\u30c8\u30ca\u30fc\u30ce\u30fc\u30c9", "pages": [ "ja/tutorials/partner-nodes/overview", "ja/tutorials/partner-nodes/faq", @@ -4515,6 +4647,12 @@ "ja/tutorials/partner-nodes/bytedance/seedream-5-lite" ] }, + { + "group": "ElevenLabs", + "pages": [ + "ja/tutorials/partner-nodes/elevenlabs/elevenlabs" + ] + }, { "group": "Google", "pages": [ @@ -4523,6 +4661,42 @@ "ja/tutorials/partner-nodes/google/nano-banana-2" ] }, + { + "group": "Grok", + "pages": [ + "ja/tutorials/partner-nodes/grok/grok" + ] + }, + { + "group": "HitPaw", + "pages": [ + "ja/tutorials/partner-nodes/hitpaw/hitpaw" + ] + }, + { + "group": "Magnific", + "pages": [ + "ja/tutorials/partner-nodes/magnific/magnific" + ] + }, + { + "group": "MiniMax", + "pages": [ + "ja/tutorials/partner-nodes/minimax/minimax" + ] + }, + { + "group": "PixVerse", + "pages": [ + "ja/tutorials/partner-nodes/pixverse/pixverse" + ] + }, + { + "group": "Quiver", + "pages": [ + "ja/tutorials/partner-nodes/quiver/quiver" + ] + }, { "group": "Stability AI", "pages": [ @@ -4588,6 +4762,12 @@ "ja/tutorials/partner-nodes/rodin/model-generation" ] }, + { + "group": "Topaz", + "pages": [ + "ja/tutorials/partner-nodes/topaz/topaz" + ] + }, { "group": "Tripo", "pages": [ @@ -4618,6 +4798,18 @@ "ja/tutorials/partner-nodes/reve/reve-image" ] }, + { + "group": "Veo", + "pages": [ + "ja/tutorials/partner-nodes/veo/veo" + ] + }, + { + "group": "Vidu", + "pages": [ + "ja/tutorials/partner-nodes/vidu/vidu" + ] + }, { "group": "Wan", "pages": [ @@ -4629,6 +4821,12 @@ "pages": [ "ja/tutorials/partner-nodes/sonilo/video-to-music" ] + }, + { + "group": "WaveSpeed", + "pages": [ + "ja/tutorials/partner-nodes/wavespeed/wavespeed" + ] } ] } @@ -4638,11 +4836,11 @@ ] }, { - "tab": "組み込みノード", + "tab": "\u7d44\u307f\u8fbc\u307f\u30ce\u30fc\u30c9", "pages": [ "ja/built-in-nodes/overview", { - "group": "ノード", + "group": "\u30ce\u30fc\u30c9", "pages": [ { "group": "3D", @@ -5125,7 +5323,7 @@ ] }, { - "group": "画像", + "group": "\u753b\u50cf", "pages": [ { "group": "BFL", @@ -5278,7 +5476,7 @@ ] }, { - "group": "ローダー", + "group": "\u30ed\u30fc\u30c0\u30fc", "pages": [ { "group": "Video Models", @@ -5308,7 +5506,7 @@ ] }, { - "group": "画像", + "group": "\u753b\u50cf", "pages": [ { "group": "Animation", @@ -5397,7 +5595,7 @@ ] }, { - "group": "アップスケーリング", + "group": "\u30a2\u30c3\u30d7\u30b9\u30b1\u30fc\u30ea\u30f3\u30b0", "pages": [ "ja/built-in-nodes/ImageScale", "ja/built-in-nodes/ImageScaleBy", @@ -5449,7 +5647,7 @@ ] }, { - "group": "ユーティリティ", + "group": "\u30e6\u30fc\u30c6\u30a3\u30ea\u30c6\u30a3", "pages": [ { "group": "Primitive", @@ -5491,7 +5689,7 @@ ] }, { - "group": "条件付け", + "group": "\u6761\u4ef6\u4ed8\u3051", "pages": [ { "group": "3D Models", @@ -5625,7 +5823,7 @@ ] }, { - "group": "モデルパッチ", + "group": "\u30e2\u30c7\u30eb\u30d1\u30c3\u30c1", "pages": [ { "group": "Chroma Radiance", @@ -5650,7 +5848,7 @@ ] }, { - "group": "潜在変数", + "group": "\u6f5c\u5728\u5909\u6570", "pages": [ { "group": "3D", @@ -5781,7 +5979,7 @@ ] }, { - "group": "サンプリング", + "group": "\u30b5\u30f3\u30d7\u30ea\u30f3\u30b0", "pages": [ { "group": "Custom Sampling", @@ -5874,7 +6072,7 @@ ] }, { - "group": "上級", + "group": "\u4e0a\u7d1a", "pages": [ { "group": "Attention Experiments", @@ -6139,7 +6337,7 @@ ] }, { - "tab": "開発", + "tab": "\u958b\u767a", "pages": [ "ja/development/overview", { @@ -6172,12 +6370,12 @@ ] }, { - "group": "カスタムノード開発", + "group": "\u30ab\u30b9\u30bf\u30e0\u30ce\u30fc\u30c9\u958b\u767a", "pages": [ "ja/custom-nodes/overview", "ja/custom-nodes/walkthrough", { - "group": "バックエンド", + "group": "\u30d0\u30c3\u30af\u30a8\u30f3\u30c9", "icon": "python", "pages": [ "ja/custom-nodes/backend/server_overview", @@ -6222,7 +6420,7 @@ ] }, { - "group": "レジストリ（Registry）", + "group": "\u30ec\u30b8\u30b9\u30c8\u30ea\uff08Registry\uff09", "pages": [ "ja/registry/overview", "ja/registry/publishing", @@ -6233,17 +6431,17 @@ ] }, { - "group": "仕様", + "group": "\u4ed5\u69d8", "pages": [ { - "group": "ワークフロー JSON", + "group": "\u30ef\u30fc\u30af\u30d5\u30ed\u30fc JSON", "pages": [ "ja/specs/workflow_json", "ja/specs/workflow_json_0.4" ] }, { - "group": "ノード定義", + "group": "\u30ce\u30fc\u30c9\u5b9a\u7fa9", "pages": [ "ja/specs/nodedef_json", "ja/specs/nodedef_json_1_0" @@ -6254,11 +6452,11 @@ ] }, { - "tab": "サポート", + "tab": "\u30b5\u30dd\u30fc\u30c8", "pages": [ "ja/support/contact-support", { - "group": "アカウント管理", + "group": "\u30a2\u30ab\u30a6\u30f3\u30c8\u7ba1\u7406", "icon": "user", "pages": [ "ja/account/create-account", @@ -6267,10 +6465,10 @@ ] }, { - "group": "請求サポート", + "group": "\u8acb\u6c42\u30b5\u30dd\u30fc\u30c8", "pages": [ { - "group": "サブスクリプション", + "group": "\u30b5\u30d6\u30b9\u30af\u30ea\u30d7\u30b7\u30e7\u30f3", "pages": [ "ja/support/subscription/subscribing", "ja/support/subscription/managing", @@ -6279,7 +6477,7 @@ ] }, { - "group": "お支払い", + "group": "\u304a\u652f\u6255\u3044", "pages": [ "ja/support/payment/accepted-payment-methods", "ja/support/payment/editing-payment-information", @@ -6292,7 +6490,7 @@ ] }, { - "group": "トラブルシューティング", + "group": "\u30c8\u30e9\u30d6\u30eb\u30b7\u30e5\u30fc\u30c6\u30a3\u30f3\u30b0", "icon": "bug", "pages": [ "ja/troubleshooting/overview", @@ -6301,7 +6499,7 @@ ] }, { - "group": "コミュニティ", + "group": "\u30b3\u30df\u30e5\u30cb\u30c6\u30a3", "pages": [ "ja/community/contributing", "ja/community/links" @@ -6310,11 +6508,11 @@ ] }, { - "tab": "Registry APIリファレンス", + "tab": "Registry API\u30ea\u30d5\u30a1\u30ec\u30f3\u30b9", "openapi": "https://api.comfy.org/openapi" }, { - "tab": "Cloud APIリファレンス", + "tab": "Cloud API\u30ea\u30d5\u30a1\u30ec\u30f3\u30b9", "openapi": { "source": "openapi-cloud.yaml", "directory": "jp/api-reference/cloud" @@ -6330,56 +6528,56 @@ }, "links": [ { - "header": "リソース", + "header": "\u30ea\u30bd\u30fc\u30b9", "items": [ { - "label": "インストール", + "label": "\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb", "href": "https://docs.comfy.org/ja/installation/system_requirements" }, { - "label": "チュートリアル", + "label": "\u30c1\u30e5\u30fc\u30c8\u30ea\u30a2\u30eb", "href": "https://docs.comfy.org/ja/tutorials/basic/text-to-image" }, { - "label": "開発", + "label": "\u958b\u767a", "href": "https://docs.comfy.org/ja/development/overview" } ] }, { - "header": "プロダクト", + "header": "\u30d7\u30ed\u30c0\u30af\u30c8", "items": [ { - "label": "機能", + "label": "\u6a5f\u80fd", "href": "https://www.comfy.org/?utm_source=docs#features-1" }, { - "label": "ギャラリー", + "label": "\u30ae\u30e3\u30e9\u30ea\u30fc", "href": "https://www.comfy.org/gallery?utm_source=docs" }, { - "label": "ダウンロード", + "label": "\u30c0\u30a6\u30f3\u30ed\u30fc\u30c9", "href": "https://www.comfy.org/download?utm_source=docs" } ] }, { - "header": "会社情報", + "header": "\u4f1a\u793e\u60c5\u5831", "items": [ { - "label": "概要", + "label": "\u6982\u8981", "href": "https://www.comfy.org/about?utm_source=docs" }, { - "label": "採用情報", + "label": "\u63a1\u7528\u60c5\u5831", "href": "https://www.comfy.org/careers?utm_source=docs" }, { - "label": "利用規約", + "label": "\u5229\u7528\u898f\u7d04", "href": "https://www.comfy.org/terms-of-service?utm_source=docs" }, { - "label": "プライバシーポリシー", + "label": "\u30d7\u30e9\u30a4\u30d0\u30b7\u30fc\u30dd\u30ea\u30b7\u30fc", "href": "https://www.comfy.org/privacy-policy?utm_source=docs" } ] @@ -6389,7 +6587,7 @@ "navbar": { "links": [ { - "label": "ダウンロード", + "label": "\u30c0\u30a6\u30f3\u30ed\u30fc\u30c9", "href": "https://comfy.org/download?utm_source=docs" } ], @@ -6599,4 +6797,4 @@ "destination": "/zh/:slug*" } ] -} +} \ No newline at end of file From be357587dee2c3e8e2bc99c00bb90c0d5c886e26 Mon Sep 17 00:00:00 2001 From: linmoumou Date: Mon, 27 Apr 2026 00:51:23 +0800 Subject: [PATCH 3/9] fix: ensure_ascii=False for docs.json to preserve Chinese/Japanese group names --- docs.json | 200 +++++++++++++++++++++++++++--------------------------- 1 file changed, 100 insertions(+), 100 deletions(-) diff --git a/docs.json b/docs.json index ad2b30e3b..7c63561c0 100644 --- a/docs.json +++ b/docs.json @@ -2190,19 +2190,19 @@ "language": "zh", "tabs": [ { - "tab": "\u5f00\u59cb\u4f7f\u7528", + "tab": "开始使用", "pages": [ { - "group": "\u5f00\u59cb\u4f7f\u7528", + "group": "开始使用", "pages": [ "zh/index", { - "group": "\u672c\u5730\u5b89\u88c5", + "group": "本地安装", "icon": "download", "pages": [ "zh/installation/system_requirements", { - "group": "Desktop(\u63a8\u8350)", + "group": "Desktop(推荐)", "pages": [ "zh/installation/desktop/windows", "zh/installation/desktop/macos", @@ -2216,7 +2216,7 @@ }, "zh/get_started/cloud", { - "group": "\u5b89\u88c5\u81ea\u5b9a\u4e49\u8282\u70b9", + "group": "安装自定义节点", "icon": "puzzle-piece", "pages": [ "zh/installation/install_custom_node", @@ -2226,7 +2226,7 @@ "zh/manager/overview", "zh/manager/install", { - "group": "\u81ea\u5b9a\u4e49\u8282\u70b9\u7ba1\u7406", + "group": "自定义节点管理", "pages": [ "zh/manager/pack-management", "zh/manager/legacy-ui" @@ -2242,7 +2242,7 @@ ] }, { - "group": "\u57fa\u7840\u6982\u5ff5", + "group": "基础概念", "pages": [ "zh/development/core-concepts/workflow", "zh/development/core-concepts/nodes", @@ -2254,7 +2254,7 @@ ] }, { - "group": "\u754c\u9762\u6307\u5357", + "group": "界面指南", "pages": [ "zh/interface/overview", "zh/interface/app-mode", @@ -2265,7 +2265,7 @@ "zh/interface/features/partial-execution", "zh/interface/features/node-docs", { - "group": "ComfyUI \u8bbe\u7f6e", + "group": "ComfyUI 设置", "icon": "gear", "pages": [ "zh/interface/settings/overview", @@ -2284,7 +2284,7 @@ ] }, { - "group": "Cloud \u4e13\u5c5e\u529f\u80fd", + "group": "Cloud 专属功能", "icon": "cloud", "pages": [ "zh/cloud/share-workflow", @@ -2294,11 +2294,11 @@ ] }, { - "group": "\u6559\u7a0b\u793a\u4f8b", + "group": "教程示例", "icon": "book", "pages": [ { - "group": "\u57fa\u7840\u793a\u4f8b", + "group": "基础示例", "pages": [ "zh/tutorials/basic/text-to-image", "zh/tutorials/basic/image-to-image", @@ -2382,7 +2382,7 @@ ] }, { - "group": "\u89c6\u9891", + "group": "视频", "pages": [ { "group": "LTX", @@ -2393,7 +2393,7 @@ ] }, { - "group": "\u4e07\u76f8\u89c6\u9891", + "group": "万相视频", "pages": [ "zh/tutorials/video/wan/wan2_2", "zh/tutorials/video/wan/wan2-2-animate", @@ -2418,7 +2418,7 @@ ] }, { - "group": "\u817e\u8baf\u6df7\u5143", + "group": "腾讯混元", "pages": [ "zh/tutorials/video/hunyuan/hunyuan-video", "zh/tutorials/video/hunyuan/hunyuan-video-1-5" @@ -2439,7 +2439,7 @@ ] }, { - "group": "\u97f3\u9891", + "group": "音频", "pages": [ { "group": "ACE-Step", @@ -2460,7 +2460,7 @@ ] }, { - "group": "\u5408\u4f5c\u4f19\u4f34\u8282\u70b9", + "group": "合作伙伴节点", "pages": [ "zh/tutorials/partner-nodes/overview", "zh/tutorials/partner-nodes/faq", @@ -2669,11 +2669,11 @@ ] }, { - "tab": "\u5185\u7f6e\u8282\u70b9", + "tab": "内置节点", "pages": [ "zh/built-in-nodes/overview", { - "group": "\u8282\u70b9", + "group": "节点", "pages": [ { "group": "3D", @@ -3156,7 +3156,7 @@ ] }, { - "group": "\u56fe\u50cf", + "group": "图像", "pages": [ { "group": "BFL", @@ -3309,7 +3309,7 @@ ] }, { - "group": "\u52a0\u8f7d\u5668", + "group": "加载器", "pages": [ { "group": "Video Models", @@ -3339,7 +3339,7 @@ ] }, { - "group": "\u56fe\u50cf", + "group": "图像", "pages": [ { "group": "Animation", @@ -3480,7 +3480,7 @@ ] }, { - "group": "\u5b9e\u7528\u5de5\u5177", + "group": "实用工具", "pages": [ { "group": "Primitive", @@ -3522,7 +3522,7 @@ ] }, { - "group": "\u6761\u4ef6", + "group": "条件", "pages": [ { "group": "3D Models", @@ -3656,7 +3656,7 @@ ] }, { - "group": "\u6a21\u578b\u8865\u4e01", + "group": "模型补丁", "pages": [ { "group": "Chroma Radiance", @@ -3681,7 +3681,7 @@ ] }, { - "group": "\u6f5c\u53d8\u91cf", + "group": "潜变量", "pages": [ { "group": "3D", @@ -3812,7 +3812,7 @@ ] }, { - "group": "\u91c7\u6837", + "group": "采样", "pages": [ { "group": "Custom Sampling", @@ -3905,7 +3905,7 @@ ] }, { - "group": "\u9ad8\u7ea7", + "group": "高级", "pages": [ { "group": "Attention Experiments", @@ -4170,7 +4170,7 @@ ] }, { - "tab": "\u5f00\u53d1", + "tab": "开发", "pages": [ "zh/development/overview", { @@ -4203,12 +4203,12 @@ ] }, { - "group": "\u5f00\u53d1\u81ea\u5b9a\u4e49\u8282\u70b9", + "group": "开发自定义节点", "pages": [ "zh/custom-nodes/overview", "zh/custom-nodes/walkthrough", { - "group": "\u540e\u7aef", + "group": "后端", "icon": "python", "pages": [ "zh/custom-nodes/backend/server_overview", @@ -4253,7 +4253,7 @@ ] }, { - "group": "\u6ce8\u518c\u8868(Registry)", + "group": "注册表(Registry)", "pages": [ "zh/registry/overview", "zh/registry/publishing", @@ -4264,17 +4264,17 @@ ] }, { - "group": "\u89c4\u8303", + "group": "规范", "pages": [ { - "group": "\u5de5\u4f5c\u6d41 JSON", + "group": "工作流 JSON", "pages": [ "zh/specs/workflow_json", "zh/specs/workflow_json_0.4" ] }, { - "group": "\u8282\u70b9\u5b9a\u4e49", + "group": "节点定义", "pages": [ "zh/specs/nodedef_json", "zh/specs/nodedef_json_1_0" @@ -4285,11 +4285,11 @@ ] }, { - "tab": "\u652f\u6301", + "tab": "支持", "pages": [ "zh/support/contact-support", { - "group": "\u8d26\u6237\u7ba1\u7406", + "group": "账户管理", "icon": "user", "pages": [ "zh/account/create-account", @@ -4298,10 +4298,10 @@ ] }, { - "group": "\u8d26\u5355\u652f\u6301", + "group": "账单支持", "pages": [ { - "group": "\u8ba2\u9605", + "group": "订阅", "pages": [ "zh/support/subscription/subscribing", "zh/support/subscription/managing", @@ -4310,7 +4310,7 @@ ] }, { - "group": "\u652f\u4ed8", + "group": "支付", "pages": [ "zh/support/payment/accepted-payment-methods", "zh/support/payment/editing-payment-information", @@ -4323,7 +4323,7 @@ ] }, { - "group": "\u6545\u969c\u6392\u9664", + "group": "故障排除", "icon": "bug", "pages": [ "zh/troubleshooting/overview", @@ -4332,7 +4332,7 @@ ] }, { - "group": "\u793e\u533a", + "group": "社区", "pages": [ "zh/community/contributing", "zh/community/links" @@ -4345,7 +4345,7 @@ "openapi": "https://api.comfy.org/openapi" }, { - "tab": "Cloud API \u53c2\u8003\u6587\u6863", + "tab": "Cloud API 参考文档", "openapi": { "source": "openapi-cloud.yaml", "directory": "zh/api-reference/cloud" @@ -4357,19 +4357,19 @@ "language": "ja", "tabs": [ { - "tab": "\u306f\u3058\u3081\u306b", + "tab": "はじめに", "pages": [ { - "group": "\u306f\u3058\u3081\u306b", + "group": "はじめに", "pages": [ "ja/index", { - "group": "\u30ed\u30fc\u30ab\u30eb\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb", + "group": "ローカルインストール", "icon": "download", "pages": [ "ja/installation/system_requirements", { - "group": "\u30c7\u30b9\u30af\u30c8\u30c3\u30d7\uff08\u63a8\u5968\uff09", + "group": "デスクトップ（推奨）", "pages": [ "ja/installation/desktop/windows", "ja/installation/desktop/macos", @@ -4383,7 +4383,7 @@ }, "ja/get_started/cloud", { - "group": "\u30ab\u30b9\u30bf\u30e0\u30ce\u30fc\u30c9\u306e\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb", + "group": "カスタムノードのインストール", "icon": "puzzle-piece", "pages": [ "ja/installation/install_custom_node", @@ -4393,7 +4393,7 @@ "ja/manager/overview", "ja/manager/install", { - "group": "\u30ab\u30b9\u30bf\u30e0\u30ce\u30fc\u30c9\u7ba1\u7406", + "group": "カスタムノード管理", "pages": [ "ja/manager/pack-management", "ja/manager/legacy-ui" @@ -4409,7 +4409,7 @@ ] }, { - "group": "\u57fa\u672c\u6982\u5ff5", + "group": "基本概念", "pages": [ "ja/development/core-concepts/workflow", "ja/development/core-concepts/nodes", @@ -4421,7 +4421,7 @@ ] }, { - "group": "\u30a4\u30f3\u30bf\u30fc\u30d5\u30a7\u30fc\u30b9\u30ac\u30a4\u30c9", + "group": "インターフェースガイド", "pages": [ "ja/interface/overview", "ja/interface/app-mode", @@ -4432,7 +4432,7 @@ "ja/interface/features/partial-execution", "ja/interface/features/node-docs", { - "group": "ComfyUI \u8a2d\u5b9a", + "group": "ComfyUI 設定", "icon": "gear", "pages": [ "ja/interface/settings/overview", @@ -4451,7 +4451,7 @@ ] }, { - "group": "Cloud \u5c02\u7528\u6a5f\u80fd", + "group": "Cloud 専用機能", "icon": "cloud", "pages": [ "ja/cloud/share-workflow", @@ -4461,11 +4461,11 @@ ] }, { - "group": "\u30c1\u30e5\u30fc\u30c8\u30ea\u30a2\u30eb", + "group": "チュートリアル", "icon": "book", "pages": [ { - "group": "\u57fa\u672c\u30c1\u30e5\u30fc\u30c8\u30ea\u30a2\u30eb", + "group": "基本チュートリアル", "pages": [ "ja/tutorials/basic/text-to-image", "ja/tutorials/basic/image-to-image", @@ -4549,7 +4549,7 @@ ] }, { - "group": "\u52d5\u753b", + "group": "動画", "pages": [ { "group": "LTX", @@ -4606,7 +4606,7 @@ ] }, { - "group": "\u30aa\u30fc\u30c7\u30a3\u30aa", + "group": "オーディオ", "pages": [ { "group": "ACE-Step", @@ -4627,7 +4627,7 @@ ] }, { - "group": "\u30d1\u30fc\u30c8\u30ca\u30fc\u30ce\u30fc\u30c9", + "group": "パートナーノード", "pages": [ "ja/tutorials/partner-nodes/overview", "ja/tutorials/partner-nodes/faq", @@ -4836,11 +4836,11 @@ ] }, { - "tab": "\u7d44\u307f\u8fbc\u307f\u30ce\u30fc\u30c9", + "tab": "組み込みノード", "pages": [ "ja/built-in-nodes/overview", { - "group": "\u30ce\u30fc\u30c9", + "group": "ノード", "pages": [ { "group": "3D", @@ -5323,7 +5323,7 @@ ] }, { - "group": "\u753b\u50cf", + "group": "画像", "pages": [ { "group": "BFL", @@ -5476,7 +5476,7 @@ ] }, { - "group": "\u30ed\u30fc\u30c0\u30fc", + "group": "ローダー", "pages": [ { "group": "Video Models", @@ -5506,7 +5506,7 @@ ] }, { - "group": "\u753b\u50cf", + "group": "画像", "pages": [ { "group": "Animation", @@ -5595,7 +5595,7 @@ ] }, { - "group": "\u30a2\u30c3\u30d7\u30b9\u30b1\u30fc\u30ea\u30f3\u30b0", + "group": "アップスケーリング", "pages": [ "ja/built-in-nodes/ImageScale", "ja/built-in-nodes/ImageScaleBy", @@ -5647,7 +5647,7 @@ ] }, { - "group": "\u30e6\u30fc\u30c6\u30a3\u30ea\u30c6\u30a3", + "group": "ユーティリティ", "pages": [ { "group": "Primitive", @@ -5689,7 +5689,7 @@ ] }, { - "group": "\u6761\u4ef6\u4ed8\u3051", + "group": "条件付け", "pages": [ { "group": "3D Models", @@ -5823,7 +5823,7 @@ ] }, { - "group": "\u30e2\u30c7\u30eb\u30d1\u30c3\u30c1", + "group": "モデルパッチ", "pages": [ { "group": "Chroma Radiance", @@ -5848,7 +5848,7 @@ ] }, { - "group": "\u6f5c\u5728\u5909\u6570", + "group": "潜在変数", "pages": [ { "group": "3D", @@ -5979,7 +5979,7 @@ ] }, { - "group": "\u30b5\u30f3\u30d7\u30ea\u30f3\u30b0", + "group": "サンプリング", "pages": [ { "group": "Custom Sampling", @@ -6072,7 +6072,7 @@ ] }, { - "group": "\u4e0a\u7d1a", + "group": "上級", "pages": [ { "group": "Attention Experiments", @@ -6337,7 +6337,7 @@ ] }, { - "tab": "\u958b\u767a", + "tab": "開発", "pages": [ "ja/development/overview", { @@ -6370,12 +6370,12 @@ ] }, { - "group": "\u30ab\u30b9\u30bf\u30e0\u30ce\u30fc\u30c9\u958b\u767a", + "group": "カスタムノード開発", "pages": [ "ja/custom-nodes/overview", "ja/custom-nodes/walkthrough", { - "group": "\u30d0\u30c3\u30af\u30a8\u30f3\u30c9", + "group": "バックエンド", "icon": "python", "pages": [ "ja/custom-nodes/backend/server_overview", @@ -6420,7 +6420,7 @@ ] }, { - "group": "\u30ec\u30b8\u30b9\u30c8\u30ea\uff08Registry\uff09", + "group": "レジストリ（Registry）", "pages": [ "ja/registry/overview", "ja/registry/publishing", @@ -6431,17 +6431,17 @@ ] }, { - "group": "\u4ed5\u69d8", + "group": "仕様", "pages": [ { - "group": "\u30ef\u30fc\u30af\u30d5\u30ed\u30fc JSON", + "group": "ワークフロー JSON", "pages": [ "ja/specs/workflow_json", "ja/specs/workflow_json_0.4" ] }, { - "group": "\u30ce\u30fc\u30c9\u5b9a\u7fa9", + "group": "ノード定義", "pages": [ "ja/specs/nodedef_json", "ja/specs/nodedef_json_1_0" @@ -6452,11 +6452,11 @@ ] }, { - "tab": "\u30b5\u30dd\u30fc\u30c8", + "tab": "サポート", "pages": [ "ja/support/contact-support", { - "group": "\u30a2\u30ab\u30a6\u30f3\u30c8\u7ba1\u7406", + "group": "アカウント管理", "icon": "user", "pages": [ "ja/account/create-account", @@ -6465,10 +6465,10 @@ ] }, { - "group": "\u8acb\u6c42\u30b5\u30dd\u30fc\u30c8", + "group": "請求サポート", "pages": [ { - "group": "\u30b5\u30d6\u30b9\u30af\u30ea\u30d7\u30b7\u30e7\u30f3", + "group": "サブスクリプション", "pages": [ "ja/support/subscription/subscribing", "ja/support/subscription/managing", @@ -6477,7 +6477,7 @@ ] }, { - "group": "\u304a\u652f\u6255\u3044", + "group": "お支払い", "pages": [ "ja/support/payment/accepted-payment-methods", "ja/support/payment/editing-payment-information", @@ -6490,7 +6490,7 @@ ] }, { - "group": "\u30c8\u30e9\u30d6\u30eb\u30b7\u30e5\u30fc\u30c6\u30a3\u30f3\u30b0", + "group": "トラブルシューティング", "icon": "bug", "pages": [ "ja/troubleshooting/overview", @@ -6499,7 +6499,7 @@ ] }, { - "group": "\u30b3\u30df\u30e5\u30cb\u30c6\u30a3", + "group": "コミュニティ", "pages": [ "ja/community/contributing", "ja/community/links" @@ -6508,11 +6508,11 @@ ] }, { - "tab": "Registry API\u30ea\u30d5\u30a1\u30ec\u30f3\u30b9", + "tab": "Registry APIリファレンス", "openapi": "https://api.comfy.org/openapi" }, { - "tab": "Cloud API\u30ea\u30d5\u30a1\u30ec\u30f3\u30b9", + "tab": "Cloud APIリファレンス", "openapi": { "source": "openapi-cloud.yaml", "directory": "jp/api-reference/cloud" @@ -6528,56 +6528,56 @@ }, "links": [ { - "header": "\u30ea\u30bd\u30fc\u30b9", + "header": "リソース", "items": [ { - "label": "\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb", + "label": "インストール", "href": "https://docs.comfy.org/ja/installation/system_requirements" }, { - "label": "\u30c1\u30e5\u30fc\u30c8\u30ea\u30a2\u30eb", + "label": "チュートリアル", "href": "https://docs.comfy.org/ja/tutorials/basic/text-to-image" }, { - "label": "\u958b\u767a", + "label": "開発", "href": "https://docs.comfy.org/ja/development/overview" } ] }, { - "header": "\u30d7\u30ed\u30c0\u30af\u30c8", + "header": "プロダクト", "items": [ { - "label": "\u6a5f\u80fd", + "label": "機能", "href": "https://www.comfy.org/?utm_source=docs#features-1" }, { - "label": "\u30ae\u30e3\u30e9\u30ea\u30fc", + "label": "ギャラリー", "href": "https://www.comfy.org/gallery?utm_source=docs" }, { - "label": "\u30c0\u30a6\u30f3\u30ed\u30fc\u30c9", + "label": "ダウンロード", "href": "https://www.comfy.org/download?utm_source=docs" } ] }, { - "header": "\u4f1a\u793e\u60c5\u5831", + "header": "会社情報", "items": [ { - "label": "\u6982\u8981", + "label": "概要", "href": "https://www.comfy.org/about?utm_source=docs" }, { - "label": "\u63a1\u7528\u60c5\u5831", + "label": "採用情報", "href": "https://www.comfy.org/careers?utm_source=docs" }, { - "label": "\u5229\u7528\u898f\u7d04", + "label": "利用規約", "href": "https://www.comfy.org/terms-of-service?utm_source=docs" }, { - "label": "\u30d7\u30e9\u30a4\u30d0\u30b7\u30fc\u30dd\u30ea\u30b7\u30fc", + "label": "プライバシーポリシー", "href": "https://www.comfy.org/privacy-policy?utm_source=docs" } ] @@ -6587,7 +6587,7 @@ "navbar": { "links": [ { - "label": "\u30c0\u30a6\u30f3\u30ed\u30fc\u30c9", + "label": "ダウンロード", "href": "https://comfy.org/download?utm_source=docs" } ], From 1ca6de16c29a670f6f1751b5ac124d535550aba9 Mon Sep 17 00:00:00 2001 From: linmoumou Date: Mon, 27 Apr 2026 00:54:00 +0800 Subject: [PATCH 4/9] docs: shorten overview paragraphs to 2-3 sentences (11 files) --- tutorials/partner-nodes/elevenlabs/elevenlabs.mdx | 2 +- tutorials/partner-nodes/grok/grok.mdx | 2 +- tutorials/partner-nodes/hitpaw/hitpaw.mdx | 10 +--------- tutorials/partner-nodes/magnific/magnific.mdx | 4 +--- tutorials/partner-nodes/minimax/minimax.mdx | 6 +----- tutorials/partner-nodes/pixverse/pixverse.mdx | 10 +--------- tutorials/partner-nodes/quiver/quiver.mdx | 12 +----------- tutorials/partner-nodes/topaz/topaz.mdx | 10 +--------- tutorials/partner-nodes/veo/veo.mdx | 8 +------- tutorials/partner-nodes/vidu/vidu.mdx | 2 +- tutorials/partner-nodes/wavespeed/wavespeed.mdx | 10 +--------- 11 files changed, 11 insertions(+), 65 deletions(-) diff --git a/tutorials/partner-nodes/elevenlabs/elevenlabs.mdx b/tutorials/partner-nodes/elevenlabs/elevenlabs.mdx index a69de824b..12256053b 100644 --- a/tutorials/partner-nodes/elevenlabs/elevenlabs.mdx +++ b/tutorials/partner-nodes/elevenlabs/elevenlabs.mdx @@ -7,7 +7,7 @@ sidebarTitle: "ElevenLabs" import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; -ElevenLabs brings world-class voice AI to ComfyUI, covering the full spectrum of audio generation and processing. **Eleven v3** delivers the most emotionally expressive speech, capturing subtle vocal nuances in intonation, pacing, and emphasis for character voiceovers and dramatic narration. **Eleven Multilingual v2** excels at stable, consistent long-form output across 29 languages, making it the choice for audiobooks and localized content. **Eleven Flash v2.5** provides near-instant speech generation with ~75ms latency at roughly half the cost, perfect for live streaming, chatbots, and real-time applications. **Scribe v2** transcribes audio to text with support for 90+ languages, speaker diarization for up to 32 speakers, 56-class entity detection, keyword prompting, and precise word-level timestamps. Beyond speech, ElevenLabs also generates sound effects from text descriptions, isolates voice from background noise, creates multi-character dialogue, and transforms one voice into another while preserving the original delivery. +ElevenLabs brings world-class voice AI to ComfyUI, covering text-to-speech, speech-to-text, sound effects, voice isolation, and voice transformation. Its **Eleven v3** and **Multilingual v2** models deliver emotionally expressive speech across 29 languages, while **Scribe v2** provides fast, accurate transcription with speaker diarization and entity detection. ## Key capabilities diff --git a/tutorials/partner-nodes/grok/grok.mdx b/tutorials/partner-nodes/grok/grok.mdx index ce893de7f..491a52381 100644 --- a/tutorials/partner-nodes/grok/grok.mdx +++ b/tutorials/partner-nodes/grok/grok.mdx @@ -7,7 +7,7 @@ sidebarTitle: "Grok" import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; -Grok Imagine by xAI brings a distinct visual style to AI image and video generation. Its **grok-imagine-image-pro** model produces the highest quality output with a signature "moody" aesthetic — dramatic lighting, filmic contrast, rich color saturation that makes every frame feel cinematic. The **grok-imagine-image** standard tier offers faster generation at ~4 seconds per image while maintaining strong visual quality, and **grok-imagine-image-beta** provides early access to experimental features. Grok excels at anime and cyberpunk aesthetics, cinematic character renders, 2D illustrations, and dynamic graphical compositions. For video, Grok supports text-to-video, video editing, video extending, and reference-to-video generation with up to 7 reference images for consistent character and scene identity across clips. +Grok Imagine by xAI brings a distinct, moody visual style to AI image and video generation. It produces cinematic-quality images with dramatic lighting and rich color saturation, and supports text-to-video, video editing, and reference-to-video generation with consistent character identity. ## Key capabilities diff --git a/tutorials/partner-nodes/hitpaw/hitpaw.mdx b/tutorials/partner-nodes/hitpaw/hitpaw.mdx index 3a396f657..23c5e20ba 100644 --- a/tutorials/partner-nodes/hitpaw/hitpaw.mdx +++ b/tutorials/partner-nodes/hitpaw/hitpaw.mdx @@ -7,15 +7,7 @@ sidebarTitle: "HitPaw" import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; -HitPaw is a well-established name in AI-powered media enhancement, known for its accessible, high-quality image and video upscaling tools. Now available as native nodes within ComfyUI, HitPaw brings its production-grade enhancement capabilities directly into your visual workflow, eliminating the need to switch between applications for post-processing. - -The ComfyUI integration covers two core workflows: **general image enhancement** (utility_hitpaw_general_image_enhance) and **video enhancement** (utility_hitpaw_video_enhance). Both are powered by HitPaw's proprietary AI models running on ComfyCloud infrastructure, meaning no local GPU resources are consumed during processing. - -The HitPaw general image enhancement model is designed as a versatile all-rounder. It handles common image quality issues including low resolution, noise, blur, and compression artifacts. Whether you're upscaling AI-generated images from Stable Diffusion or Midjourney, restoring old photographs, or preparing product images for e-commerce, HitPaw produces clean, natural-looking results without the over-sharpened or "plastic" appearance that some upscalers introduce. The model intelligently distinguishes between different types of image content — portraits, landscapes, architecture, text — and applies enhancement strategies optimized for each category. - -On the video side, HitPaw video enhancement performs frame-by-frame AI upscaling and denoising. It is particularly effective for improving the quality of compressed video footage, screen recordings, and low-resolution source material. The model maintains temporal consistency across frames, preventing the flickering or "pumping" artifacts that can occur with naive frame-independent processing. This makes it suitable for everything from archival video restoration to improving the quality of AI-generated video clips before final export. - -Both workflows are designed to be straightforward drop-in enhancements. You can connect them to any image or video output node in ComfyUI — whether from generative models, renders, or imported media — and get consistently good results without extensive parameter tuning. The HitPaw nodes are an excellent choice for creators who want reliable, predictable enhancement quality with minimal configuration overhead. +HitPaw brings production-grade AI image and video enhancement to ComfyUI as native nodes. Its **general image enhancement** and **video enhancement** models handle upscaling, denoising, and restoration with minimal configuration, all powered by ComfyCloud with no local GPU needed. ## Key capabilities diff --git a/tutorials/partner-nodes/magnific/magnific.mdx b/tutorials/partner-nodes/magnific/magnific.mdx index 8343bb839..0b72372a7 100644 --- a/tutorials/partner-nodes/magnific/magnific.mdx +++ b/tutorials/partner-nodes/magnific/magnific.mdx @@ -7,9 +7,7 @@ sidebarTitle: "Magnific AI" import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; -Magnific AI is a cutting-edge AI image enhancement platform that specializes in high-quality upscaling, creative image transformation, and intelligent photo retouching. Now natively integrated as partner nodes in ComfyUI, Magnific brings professional-grade image enhancement capabilities directly into your workflows. - -The Magnific node suite offers five distinct modes to handle a wide range of image enhancement tasks. **Precise Upscale** delivers pixel-perfect enlargement with minimal deviation from the original, making it ideal for preserving fidelity in product shots or archival photos. **Creative Upscale**, in contrast, uses AI-driven semantic interpretation to add realistic detail at up to 16x amplification — perfect for breathing new life into low-resolution images. The **Creativity slider** in Creative mode gives you fine-grained control over how much new detail the model introduces, from subtle texture enhancement to bold reinterpretation. **Style Transfer** applies a reference image's aesthetic — color palette, lighting mood, brushstroke texture — onto your source image, enabling artistic transformations without manual masking. **Relight** lets you control the lighting direction, intensity, and color temperature, making it easy to match subject lighting with a target environment. **Skin Enhancer** provides intelligent portrait retouching by smoothing skin texture while preserving natural features, reducing blemishes and evening out skin tone automatically. +Magnific AI is a professional-grade image enhancement platform, now natively integrated into ComfyUI. It offers **Precise** and **Creative** upscaling (up to 16x), style transfer, relighting, and skin enhancement — bringing studio-quality post-processing directly into your workflow. diff --git a/tutorials/partner-nodes/minimax/minimax.mdx b/tutorials/partner-nodes/minimax/minimax.mdx index 3730ec3af..dcfebefcc 100644 --- a/tutorials/partner-nodes/minimax/minimax.mdx +++ b/tutorials/partner-nodes/minimax/minimax.mdx @@ -7,11 +7,7 @@ sidebarTitle: "MiniMax (Hailuo)" import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; -MiniMax, also known as Hailuo, is a leading AI video generation platform recognized for its exceptional motion quality and cinematic output. Now available inside ComfyUI, MiniMax offers three core model families that cover a comprehensive range of video generation use cases — from cinematic narrative production to character-driven content creation. - -T2V-01 is the text-to-video model that supports the Director mode with 15 distinct camera movement types: pan, tilt, dolly, crane, tracking, aerial, handheld, roll, zoom-in, zoom-out, push-in, pull-out, orbit, first-person, and static. This gives you granular control over shot composition and cinematic storytelling without leaving the ComfyUI node graph. I2V-01 takes image-to-video generation further with two sub-modes: I2V-01-Director for stylized camera direction with image input, and I2V-01-Live for generating dynamic, lively motion from a single image, optimized for natural movement like flowing hair, swaying grass, and gentle camera breathing. S2V-01 introduces subject reference to video, allowing you to supply a reference image of a specific character or object and maintain its identity throughout the generated clip — essential for branded content, serial characters, and product consistency. - -All MiniMax models are known for industry-leading motion quality, with realistic physics in character movement, fluid dynamics, and environmental effects. The Director mode is particularly powerful for content creators who need precise camera language to tell their stories — each of the 15 movement types produces a distinctly different cinematic feel, from the intimacy of handheld to the grandeur of aerial crane shots. +MiniMax (Hailuo) is a leading AI video generation platform renowned for exceptional motion quality and cinematic output, now inside ComfyUI. Its **T2V-01** (text-to-video with 15 camera directions), **I2V-01** (image-to-video with live motion), and **S2V-01** (subject reference) models cover everything from film-grade narrative to character-driven content. ## Key capabilities diff --git a/tutorials/partner-nodes/pixverse/pixverse.mdx b/tutorials/partner-nodes/pixverse/pixverse.mdx index ec26df97b..30100f74d 100644 --- a/tutorials/partner-nodes/pixverse/pixverse.mdx +++ b/tutorials/partner-nodes/pixverse/pixverse.mdx @@ -7,15 +7,7 @@ sidebarTitle: "PixVerse AI" import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; -PixVerse is a state-of-the-art AI video generation platform that produces high-quality, cinematic videos from text descriptions and reference images. Now integrated as partner nodes in ComfyUI, PixVerse brings its powerful video generation capabilities directly into your workflows without requiring a separate API key or external setup. - -The PixVerse node suite currently supports three generation modes. **Text-to-Video** allows you to create original videos from nothing but a text prompt — describe a scene, a character, an action, or a cinematic mood, and PixVerse generates a corresponding video clip. This is ideal for concept visualization, storyboarding, and rapid ideation. **Image-to-Video** takes a reference image as input and animates it, preserving the subject and scene composition while introducing motion. This mode excels at bringing static artwork, product shots, or photographs to life with natural movement. **Template Image-to-Video** provides a more structured approach: it uses a pre-defined motion template applied to your input image, giving you predictable and repeatable animation styles without the variability of open-ended generation. This is especially useful for production pipelines where consistent motion characteristics are required across multiple assets. - -All three modes share a common set of parameters for fine-tuning the output, including resolution, duration, and style guidance. The generated videos can be combined with other ComfyUI nodes for further processing — overlay text, composite with other clips, or pipe into a video-to-video enhancement workflow. - -PixVerse is particularly well-suited for content creators, marketers, and artists who need rapid video prototyping without the overhead of traditional animation pipelines. For example, a marketer can use Text-to-Video to generate product demonstration clips from copy alone, while a digital artist can use Image-to-Video to breathe motion into their illustrations with a single click. The Template I2V mode provides an additional layer of consistency: when generating video variants for A/B testing or batch processing, the fixed motion templates ensure that differences in output are driven by the input image rather than random motion seeds, making comparison and iteration more reliable. - -The videos generated by PixVerse can be further processed in ComfyUI by chaining with other nodes. Common post-processing workflows include applying video filters via VHS nodes, compositing PixVerse output over a background video, extracting frames for frame-by-frame editing, or using audio-reactive nodes to sync generated video with a soundtrack. This composability makes PixVerse a versatile building block rather than an isolated tool — it fits naturally into larger production pipelines spanning concept development, asset creation, and final rendering. +PixVerse is an AI video generation platform that produces cinematic videos from text and images, now integrated as partner nodes in ComfyUI. It supports **Text-to-Video**, **Image-to-Video**, and **Template Image-to-Video** modes for rapid, repeatable video creation without separate API keys or external setup. diff --git a/tutorials/partner-nodes/quiver/quiver.mdx b/tutorials/partner-nodes/quiver/quiver.mdx index 8817d2180..fb348d10d 100644 --- a/tutorials/partner-nodes/quiver/quiver.mdx +++ b/tutorials/partner-nodes/quiver/quiver.mdx @@ -7,17 +7,7 @@ sidebarTitle: "Quiver AI" import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; -Quiver is an AI-powered vector graphics platform that specializes in generating high-quality SVG (Scalable Vector Graphics) from text descriptions and reference images. Now integrated as partner nodes in ComfyUI, Quiver enables you to create clean, scalable vector artwork directly within your workflows — no manual vector editing required. - -The Quiver node suite offers two generation modes. **Text-to-SVG** lets you describe the vector graphic you want in natural language, and Quiver generates a fully editable SVG file. This is ideal for creating icons, illustrations, logos, diagrams, and UI elements without needing graphic design skills. Just describe what you want — "a minimalist mountain landscape silhouette," "a three-icon set for settings, profile, and notifications" — and the SVG is generated ready for use in web design, print, or further editing in tools like Figma or Illustrator. **Image-to-SVG** takes an existing raster image (PNG, JPG, etc.) as input and converts it into a clean vector representation. This process traces the shapes, colors, and gradients from the source image and reconstructs them as scalable vector paths, making it perfect for converting hand-drawn sketches, bitmap logos, or photographed artwork into resolution-independent SVGs. - -Quiver supports three model variants that offer different trade-offs between generation quality and speed. **Arrow 1.1** is the standard model — fast and efficient, suitable for most use cases. **Arrow 1.1 Max** is the high-quality variant that produces more detailed and accurate vector output, ideal for complex illustrations where visual fidelity matters most. **Arrow Preview** offers early access to experimental features and upcoming model improvements. Each variant can be selected directly from the node's model parameter, giving you the flexibility to choose the right balance of speed and quality for your specific task. - -Vector graphics generated by Quiver have significant advantages over raster images in many design contexts. SVGs are resolution-independent — they look sharp at any display size, from mobile icons to billboards. They also have smaller file sizes for geometric artwork, can be styled with CSS, animated with SMIL or JavaScript, and embedded directly into web pages without additional HTTP requests. This makes Quiver's output immediately useful for web development, UI/UX design, screen printing, laser engraving, and any application where scalable resolution and small file footprints matter. - -In practice, Quiver integrates naturally with typical design workflows in ComfyUI. For Text-to-SVG, try prompting with specific style cues — "flat design icon," "line art illustration," "filled silhouette" — to guide the model toward the desired vector aesthetic. For Image-to-SVG, best results come from source images with clear shapes, high contrast, and limited color palettes; complex photographic images with many gradients may produce more path segments and larger file sizes. The Arrow 1.1 Max model is recommended when the SVG will be scaled up significantly (e.g., for large-format printing), while Arrow 1.1 is sufficient for web assets and icons where speed is a priority. - -Once generated, SVGs can be output through ComfyUI's standard save nodes or piped into downstream processing. You can combine Quiver-generated SVGs with raster nodes — for example, rasterizing an SVG at a specific resolution and feeding it into a GPT-Image-1 or Magnific enhancement node — creating hybrid workflows that leverage the strengths of both vector precision and AI image generation. +Quiver is an AI-powered vector graphics platform that generates high-quality SVGs from text descriptions and raster images, now available as partner nodes in ComfyUI. With models from **Arrow 1.1** (fast) to **Arrow 1.1 Max** (high-fidelity), it produces resolution-independent vector artwork without manual editing. diff --git a/tutorials/partner-nodes/topaz/topaz.mdx b/tutorials/partner-nodes/topaz/topaz.mdx index f923550d3..f9b21f2d7 100644 --- a/tutorials/partner-nodes/topaz/topaz.mdx +++ b/tutorials/partner-nodes/topaz/topaz.mdx @@ -7,15 +7,7 @@ sidebarTitle: "Topaz Labs" import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; -Topaz Labs has long been the gold standard in AI-powered image and video enhancement, trusted by professional photographers, videographers, and post-production studios worldwide. Now integrated into ComfyUI, Topaz's suite of models brings enterprise-grade upscaling, restoration, and enhancement directly into your visual workflow — no more round-tripping between applications. - -The integration covers two primary domains: **image enhancement** and **video enhancement**. On the image side, **api_topaz_image_enhance** provides general-purpose AI upscaling and denoising, while the **Starlight** diffusion-based upscaler (utility_topaz_landscape_upscaler) specializes in landscape photography — it uses a diffusion process to hallucinate realistic detail beyond simple pixel interpolation, producing stunning 8K-level output from lower-resolution sources. The **Reimagine** mode takes this further by allowing creative reinterpretation: you can guide the upscaling with a text prompt, transforming a simple photo into a completely restyled or enhanced version. For illustration and anime art, **utility_topaz_illustration_upscale** applies models optimized for flat colors, line art, and non-photographic content, preserving crisp edges and smooth gradients. - -On the video side, **api_topaz_video_enhance** brings Topaz's video AI to ComfyUI, enabling frame-by-frame intelligent upscaling, denoising, deinterlacing, and stabilization. The **Apollo** frame interpolation model (included in the video pipeline) generates intermediate frames to create smooth slow-motion or increase frame rates, making it ideal for converting 24fps footage to 60fps or beyond. - -One standout feature is **face enhancement** — Topaz's specialized facial detail reconstruction model that detects and enhances faces within both images and video frames. This is particularly valuable for portrait photography, interview footage, and any content where human faces are the focus, as traditional upscaling often leaves facial features blurry or unnatural. - -All Topaz nodes in ComfyUI are powered by the Topaz API via ComfyCloud, meaning no local GPU is required for computation. You can chain Topaz enhancement with other ComfyUI nodes for a complete pipeline — generate with AI, enhance with Topaz, and composite in a single session. +Topaz Labs brings enterprise-grade AI image and video enhancement to ComfyUI — trusted by professionals worldwide. Its suite covers **upscaling**, **denoising**, **face enhancement**, **Starlight diffusion upscale** for landscapes, **Reimagine** creative upscaling, and the **Apollo** frame interpolation model for video. ## Key capabilities diff --git a/tutorials/partner-nodes/veo/veo.mdx b/tutorials/partner-nodes/veo/veo.mdx index faee76908..53a562042 100644 --- a/tutorials/partner-nodes/veo/veo.mdx +++ b/tutorials/partner-nodes/veo/veo.mdx @@ -7,13 +7,7 @@ sidebarTitle: "Veo" import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; -Google DeepMind's Veo family represents the state of the art in AI video generation, now deeply integrated into ComfyUI through a suite of dedicated workflow nodes. The flagship model, **Veo 3.1**, delivers cinema-grade 4K video with native audio — a breakthrough that brings full-production-quality video generation into a single pipeline. Whether you're creating short films, product demos, social media content, or concept visualizations, Veo 3.1 produces realistic motion, consistent character identities, and scene-coherent narratives that rival traditional film production. - -Alongside Veo 3.1, **Veo 2.0** offers a lighter, faster text-to-video and image-to-video alternative that balances quality with speed, making it ideal for rapid prototyping and iterative creative workflows. Veo 2.0's image-to-video capability (api_veo2_i2v) lets you animate still images with natural motion, breathing life into concept art, photography, or AI-generated visuals. - -All Veo nodes are powered by Google's infrastructure and accessed via the ComfyCloud platform. The tight integration means you can chain Veo generation with ComfyUI's post-processing nodes — upscaling, color grading, audio mixing, and compositing — all in a single visual workflow. This eliminates the need to export and re-import media between different tools, dramatically accelerating the video production pipeline. - -Key differentiators of Veo in ComfyUI include: native audio generation synchronized with video, high-fidelity 4K output that preserves fine details and textures, temporal consistency across frames that eliminates flickering and jarring transitions, and support for detailed prompt-guided cinematography including camera angles, lighting conditions, and scene composition. For creators who need production-ready video from text or image prompts, Veo represents the most advanced AI video generation option available today. +Google DeepMind's **Veo 3.1** and **Veo 2.0** bring state-of-the-art AI video generation to ComfyUI. Veo 3.1 delivers cinema-grade 4K video with native audio, while Veo 2.0 offers fast text-to-video and image-to-video for rapid prototyping — all seamlessly integrated into your ComfyUI pipeline. ## Key capabilities diff --git a/tutorials/partner-nodes/vidu/vidu.mdx b/tutorials/partner-nodes/vidu/vidu.mdx index d8f58f3b9..3fc55f1b2 100644 --- a/tutorials/partner-nodes/vidu/vidu.mdx +++ b/tutorials/partner-nodes/vidu/vidu.mdx @@ -7,7 +7,7 @@ sidebarTitle: "Vidu" import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; -Vidu is an AI video generation platform developed by Shengshu Technology, known for its fast inference speed and high visual quality across multiple model generations. Now available inside ComfyUI, Vidu offers three model families — Q1, Q2, and Q3 — each tailored to different creative needs. Q1 provides the core foundation for text-to-video, image-to-video, reference-to-video, start-end-to-video, and video extension workflows. Q2 brings 3x faster generation, support for up to 7 reference images, stronger subject consistency, and cinematic camera language controls including pan, zoom, and rotation. Q3 is the latest generation, delivering the highest visual fidelity with exceptional performance in anime and 2D animation styles, and can generate a 4-second video clip in as fast as 10 seconds. Whether you need quick drafts, polished character-driven narratives, or stylized animation sequences, Vidu's tiered model lineup gives you the flexibility to balance speed, quality, and creative control. +Vidu is a fast AI video generation platform by Shengshu Technology, now available inside ComfyUI. Its **Q1**, **Q2**, and **Q3** model families deliver a range of speed and quality options — from foundation generation to 3x faster inference with camera control and exceptional anime output. ## Key capabilities diff --git a/tutorials/partner-nodes/wavespeed/wavespeed.mdx b/tutorials/partner-nodes/wavespeed/wavespeed.mdx index 9611c2aad..322bba19e 100644 --- a/tutorials/partner-nodes/wavespeed/wavespeed.mdx +++ b/tutorials/partner-nodes/wavespeed/wavespeed.mdx @@ -7,15 +7,7 @@ sidebarTitle: "WaveSpeed" import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; -WaveSpeed is an emerging AI media enhancement platform that specializes in high-performance upscaling and restoration models, now available as native workflow nodes in ComfyUI. The integration brings three distinct capabilities to your toolbox: **FLSH-VSR video upscaling**, **general image upscaling**, and the **SeedVR2 AI image fix** for restoring damaged or low-quality photographs. All processing is handled via the ComfyCloud API, requiring no local GPU resources. - -**FLSH-VSR** (api_wavespeed_flshvsr_video_upscale) is WaveSpeed's flagship video super-resolution model. It takes low-resolution video input and intelligently reconstructs high-resolution output frame by frame, with particular strength in temporal coherence — meaning the enhanced frames maintain consistent detail and lighting across cuts and movements. Unlike simpler frame-independent upscalers that can introduce flickering or jitter, FLSH-VSR analyzes motion vectors between frames to produce smooth, realistic results. This makes it ideal for upscaling archived video footage, improving AI-generated video quality, or converting SD content to HD or beyond. - -For standalone image upscaling, **api_wavespped_image_upscale** (note the typo in the template identifier, preserved for consistency) provides a robust general-purpose model. It handles common upscaling scenarios — enlarging AI-generated images, improving photograph resolution, and cleaning up compression artifacts — with natural-looking detail reconstruction. The model works across diverse content types including photography, digital art, renders, and screenshots. - -The **SeedVR2 AI image fix** (api_wavespped_seedvr2_ai_image_fix) is a specialized restoration model designed for recovering damaged, degraded, or extremely low-quality images. It goes beyond simple upscaling by performing intelligent reconstruction of missing or corrupted regions: it can remove noise patterns, repair compression blocking, fix color degradation, and reconstruct lost facial details in old photographs. This makes it an excellent choice for photo restoration projects, recovering detail from heavily compressed web images, or salvaging generations that turned out blurry or artifact-ridden. SeedVR2 applies a diffusion-based approach to "imagine" the most likely clean version of the input, producing results that look natural rather than artificially sharpened. - -All three WaveSpeed nodes are designed as straightforward single-pass operations that integrate easily into larger ComfyUI pipelines. Whether you're enhancing final output or preprocessing source material for further generation, WaveSpeed provides reliable, high-quality results with minimal configuration. +WaveSpeed is a high-performance AI media enhancement platform, now available as native workflow nodes in ComfyUI. It offers **FLSH-VSR video upscaling**, **general image upscaling**, and the **SeedVR2 AI image fix** for restoring damaged photos — all powered by ComfyCloud with no local GPU needed. ## Key capabilities From ae557df8596cbd61534e1c9426757297d3de70e9 Mon Sep 17 00:00:00 2001 From: linmoumou Date: Mon, 27 Apr 2026 01:36:55 +0800 Subject: [PATCH 5/9] docs: restructure 11 P0 partner-node pages into hierarchical format Split each single provider .mdx into overview + per-model subpages: - ElevenLabs: overview + 6 model pages (multilingual-v2, scribe-v2, sts, sfx, dialogue, isolation) - Grok: overview + grok-imagine + grok-video - Vidu: overview + q1/q2/q3 - MiniMax: overview + minimax-01 + minimax-s2v-01 - Magnific: overview + upscale + creative-tools - PixVerse: overview + video - Quiver: overview + svg - Veo: overview + veo-2-0 + veo-3-1 - Topaz: overview + image + video-enhance - HitPaw: overview + image-enhance + video-enhance - WaveSpeed: overview + upscale + flashvsr New structure: overview.mdx (platform intro + model index) + per-model .mdx (capabilities + workflows) --- .../elevenlabs/eleven-multilingual-v2.mdx | 31 ++++ .../partner-nodes/elevenlabs/elevenlabs.mdx | 102 ------------- .../partner-nodes/elevenlabs/overview.mdx | 47 ++++++ .../partner-nodes/elevenlabs/scribe-v2.mdx | 31 ++++ .../elevenlabs/speech-to-speech.mdx | 30 ++++ .../elevenlabs/text-to-dialogue.mdx | 30 ++++ .../elevenlabs/text-to-sound-effects.mdx | 30 ++++ .../elevenlabs/voice-isolation.mdx | 31 ++++ tutorials/partner-nodes/grok/grok-imagine.mdx | 47 ++++++ .../grok/{grok.mdx => grok-video.mdx} | 44 ++---- tutorials/partner-nodes/grok/overview.mdx | 35 +++++ .../hitpaw/hitpaw-image-enhance.mdx | 27 ++++ .../hitpaw/hitpaw-video-enhance.mdx | 27 ++++ .../hitpaw/{hitpaw.mdx => overview.mdx} | 25 ++-- ...gnific.mdx => magnific-creative-tools.mdx} | 36 +---- .../magnific/magnific-upscale.mdx | 46 ++++++ tutorials/partner-nodes/magnific/overview.mdx | 32 +++++ .../partner-nodes/minimax/minimax-01.mdx | 48 +++++++ .../partner-nodes/minimax/minimax-s2v-01.mdx | 31 ++++ tutorials/partner-nodes/minimax/minimax.mdx | 63 -------- tutorials/partner-nodes/minimax/overview.mdx | 32 +++++ tutorials/partner-nodes/pixverse/overview.mdx | 27 ++++ .../{pixverse.mdx => pixverse-video.mdx} | 8 +- tutorials/partner-nodes/quiver/overview.mdx | 29 ++++ .../quiver/{quiver.mdx => quiver-svg.mdx} | 11 +- tutorials/partner-nodes/topaz/overview.mdx | 36 +++++ tutorials/partner-nodes/topaz/topaz-image.mdx | 50 +++++++ .../topaz/topaz-video-enhance.mdx | 28 ++++ tutorials/partner-nodes/topaz/topaz.mdx | 61 -------- tutorials/partner-nodes/veo/overview.mdx | 35 +++++ tutorials/partner-nodes/veo/veo-2-0.mdx | 26 ++++ tutorials/partner-nodes/veo/veo-3-1.mdx | 40 ++++++ tutorials/partner-nodes/veo/veo.mdx | 51 ------- tutorials/partner-nodes/vidu/overview.mdx | 36 +++++ tutorials/partner-nodes/vidu/vidu-q1.mdx | 77 ++++++++++ tutorials/partner-nodes/vidu/vidu-q2.mdx | 66 +++++++++ tutorials/partner-nodes/vidu/vidu-q3.mdx | 44 ++++++ tutorials/partner-nodes/vidu/vidu.mdx | 135 ------------------ .../partner-nodes/wavespeed/overview.mdx | 35 +++++ .../wavespeed/wavespeed-flashvsr.mdx | 27 ++++ .../wavespeed/wavespeed-upscale.mdx | 40 ++++++ .../partner-nodes/wavespeed/wavespeed.mdx | 51 ------- 42 files changed, 1180 insertions(+), 558 deletions(-) create mode 100644 tutorials/partner-nodes/elevenlabs/eleven-multilingual-v2.mdx delete mode 100644 tutorials/partner-nodes/elevenlabs/elevenlabs.mdx create mode 100644 tutorials/partner-nodes/elevenlabs/overview.mdx create mode 100644 tutorials/partner-nodes/elevenlabs/scribe-v2.mdx create mode 100644 tutorials/partner-nodes/elevenlabs/speech-to-speech.mdx create mode 100644 tutorials/partner-nodes/elevenlabs/text-to-dialogue.mdx create mode 100644 tutorials/partner-nodes/elevenlabs/text-to-sound-effects.mdx create mode 100644 tutorials/partner-nodes/elevenlabs/voice-isolation.mdx create mode 100644 tutorials/partner-nodes/grok/grok-imagine.mdx rename tutorials/partner-nodes/grok/{grok.mdx => grok-video.mdx} (53%) create mode 100644 tutorials/partner-nodes/grok/overview.mdx create mode 100644 tutorials/partner-nodes/hitpaw/hitpaw-image-enhance.mdx create mode 100644 tutorials/partner-nodes/hitpaw/hitpaw-video-enhance.mdx rename tutorials/partner-nodes/hitpaw/{hitpaw.mdx => overview.mdx} (51%) rename tutorials/partner-nodes/magnific/{magnific.mdx => magnific-creative-tools.mdx} (52%) create mode 100644 tutorials/partner-nodes/magnific/magnific-upscale.mdx create mode 100644 tutorials/partner-nodes/magnific/overview.mdx create mode 100644 tutorials/partner-nodes/minimax/minimax-01.mdx create mode 100644 tutorials/partner-nodes/minimax/minimax-s2v-01.mdx delete mode 100644 tutorials/partner-nodes/minimax/minimax.mdx create mode 100644 tutorials/partner-nodes/minimax/overview.mdx create mode 100644 tutorials/partner-nodes/pixverse/overview.mdx rename tutorials/partner-nodes/pixverse/{pixverse.mdx => pixverse-video.mdx} (81%) create mode 100644 tutorials/partner-nodes/quiver/overview.mdx rename tutorials/partner-nodes/quiver/{quiver.mdx => quiver-svg.mdx} (66%) create mode 100644 tutorials/partner-nodes/topaz/overview.mdx create mode 100644 tutorials/partner-nodes/topaz/topaz-image.mdx create mode 100644 tutorials/partner-nodes/topaz/topaz-video-enhance.mdx delete mode 100644 tutorials/partner-nodes/topaz/topaz.mdx create mode 100644 tutorials/partner-nodes/veo/overview.mdx create mode 100644 tutorials/partner-nodes/veo/veo-2-0.mdx create mode 100644 tutorials/partner-nodes/veo/veo-3-1.mdx delete mode 100644 tutorials/partner-nodes/veo/veo.mdx create mode 100644 tutorials/partner-nodes/vidu/overview.mdx create mode 100644 tutorials/partner-nodes/vidu/vidu-q1.mdx create mode 100644 tutorials/partner-nodes/vidu/vidu-q2.mdx create mode 100644 tutorials/partner-nodes/vidu/vidu-q3.mdx delete mode 100644 tutorials/partner-nodes/vidu/vidu.mdx create mode 100644 tutorials/partner-nodes/wavespeed/overview.mdx create mode 100644 tutorials/partner-nodes/wavespeed/wavespeed-flashvsr.mdx create mode 100644 tutorials/partner-nodes/wavespeed/wavespeed-upscale.mdx delete mode 100644 tutorials/partner-nodes/wavespeed/wavespeed.mdx diff --git a/tutorials/partner-nodes/elevenlabs/eleven-multilingual-v2.mdx b/tutorials/partner-nodes/elevenlabs/eleven-multilingual-v2.mdx new file mode 100644 index 000000000..37c72148f --- /dev/null +++ b/tutorials/partner-nodes/elevenlabs/eleven-multilingual-v2.mdx @@ -0,0 +1,31 @@ +--- +title: "ElevenLabs Text-to-Speech — Multilingual v2 / v3 / Flash v2.5" +description: "Generate lifelike, emotionally expressive speech from text using ElevenLabs' TTS models in ComfyUI." +sidebarTitle: "Eleven Multilingual v2 / v3" +--- + +import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; +import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; + +Generate speech from text using ElevenLabs' flagship Text-to-Speech models. Choose between the high-quality **Eleven v3**, the stable **Multilingual v2** supporting 29 languages, or the low-latency **Flash v2.5** for performance-critical applications. + +## Key capabilities + +- **Emotionally expressive speech** — Natural intonation, pacing, and tone across all three model tiers +- **Multi-language support** — 29 languages with Multilingual v2, broader coverage with Flash v2.5 +- **Low-latency option** — Flash v2.5 for real-time and performance-sensitive use cases +- **Voice customization** — Use ElevenLabs voice library or clone custom voices + + + + +## Available workflows + + + + Generate speech from text directly in your browser. + + + Download the workflow JSON. + + diff --git a/tutorials/partner-nodes/elevenlabs/elevenlabs.mdx b/tutorials/partner-nodes/elevenlabs/elevenlabs.mdx deleted file mode 100644 index 12256053b..000000000 --- a/tutorials/partner-nodes/elevenlabs/elevenlabs.mdx +++ /dev/null @@ -1,102 +0,0 @@ ---- -title: "ElevenLabs Text-to-Speech, Speech-to-Text, Sound Effects & Voice Tools" -description: "Generate natural, emotionally rich speech, transcribe audio, create sound effects, and isolate voices using ElevenLabs in ComfyUI." -sidebarTitle: "ElevenLabs" ---- - -import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; -import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; - -ElevenLabs brings world-class voice AI to ComfyUI, covering text-to-speech, speech-to-text, sound effects, voice isolation, and voice transformation. Its **Eleven v3** and **Multilingual v2** models deliver emotionally expressive speech across 29 languages, while **Scribe v2** provides fast, accurate transcription with speaker diarization and entity detection. - -## Key capabilities - -- **Text to Speech** — Generate lifelike audio from text with three model tiers (Eleven v3, Multilingual v2, Flash v2.5) -- **Speech to Speech** — Transform a source audio clip into a different voice while keeping the original intonation and pacing -- **Speech to Text** — Transcribe audio with Scribe v2: 90+ languages, 32-speaker diarization, entity detection, keyword boost -- **Sound Effects** — Generate audio effects from text descriptions -- **Voice Isolation** — Separate voice from background noise, music, or ambient sound -- **Multi-Character Dialogue** — Generate conversations between multiple distinct voices in a single pass - - - - -## Available workflows - -### Text to Speech - -Generate speech from text using the flagship Eleven v3, the stable Multilingual v2, or the low-latency Flash v2.5 model. - - - - Generate speech from text directly in your browser. - - - Download the workflow JSON. - - - -### Speech to Speech - -Take an existing audio recording and transform it into a different voice or style while preserving the spoken content and delivery. - - - - Transform existing audio into a new voice. - - - Download the workflow JSON. - - - -### Speech to Text - -Transcribe audio to text using Scribe v2, with advanced speaker diarization and entity detection. - - - - Transcribe audio to text directly in your workflow. - - - Download the workflow JSON. - - - -### Text to Sound Effects - -Generate sound effects, ambient audio, and foley from text descriptions. - - - - Generate sound effects from text. - - - Download the workflow JSON. - - - -### Voice Isolation - -Separate the voice track from background noise, music, or ambient sound in any audio recording. - - - - Isolate voice from noisy audio. - - - Download the workflow JSON. - - - -### Text to Dialogue - -Generate multi-character dialogue with distinct voices from a single text input. - - - - Generate multi-voice dialogue from text. - - - Download the workflow JSON. - - diff --git a/tutorials/partner-nodes/elevenlabs/overview.mdx b/tutorials/partner-nodes/elevenlabs/overview.mdx new file mode 100644 index 000000000..9e7f2450b --- /dev/null +++ b/tutorials/partner-nodes/elevenlabs/overview.mdx @@ -0,0 +1,47 @@ +--- +title: "ElevenLabs Text-to-Speech, Speech-to-Text, Sound Effects & Voice Tools" +description: "Generate natural, emotionally rich speech, transcribe audio, create sound effects, and isolate voices using ElevenLabs in ComfyUI." +sidebarTitle: "ElevenLabs" +--- + +import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; +import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; + +ElevenLabs brings world-class voice AI to ComfyUI, covering text-to-speech, speech-to-text, sound effects, voice isolation, and voice transformation. Its **Eleven v3** and **Multilingual v2** models deliver emotionally expressive speech across 29 languages, while **Scribe v2** provides fast, accurate transcription with speaker diarization and entity detection. + +## Key capabilities + +- **Text to Speech** — Generate lifelike audio from text with three model tiers (Eleven v3, Multilingual v2, Flash v2.5) +- **Speech to Speech** — Transform a source audio clip into a different voice while keeping the original intonation and pacing +- **Speech to Text** — Transcribe audio with Scribe v2: 90+ languages, 32-speaker diarization, entity detection, keyword boost +- **Sound Effects** — Generate audio effects from text descriptions +- **Voice Isolation** — Separate voice from background noise, music, or ambient sound +- **Multi-Character Dialogue** — Generate conversations between multiple distinct voices in a single pass + + + + +## Models + +ElevenLabs offers a suite of specialized voice AI models, each optimized for a distinct task. From high-fidelity speech generation to real-time transcription and audio editing, the following models are available as workflow-powered nodes in ComfyUI. + + + + High-fidelity text-to-speech with three model tiers. + + + Fast, accurate speech-to-text with speaker diarization. + + + Transform audio into a different voice while preserving delivery. + + + Generate sound effects and ambient audio from text. + + + Multi-character dialogue with distinct voices in one pass. + + + Separate voice from background noise and music. + + diff --git a/tutorials/partner-nodes/elevenlabs/scribe-v2.mdx b/tutorials/partner-nodes/elevenlabs/scribe-v2.mdx new file mode 100644 index 000000000..879539cf3 --- /dev/null +++ b/tutorials/partner-nodes/elevenlabs/scribe-v2.mdx @@ -0,0 +1,31 @@ +--- +title: "ElevenLabs Speech-to-Text — Scribe v2" +description: "Transcribe audio to text with high accuracy, speaker diarization, and entity detection using ElevenLabs Scribe v2 in ComfyUI." +sidebarTitle: "Scribe v2" +--- + +import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; +import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; + +Transcribe audio to text using **Scribe v2**, ElevenLabs' advanced speech-to-text model. It supports over 90 languages with speaker diarization for up to 32 speakers, entity detection, and keyword boosting for domain-specific terminology. + +## Key capabilities + +- **90+ language support** — Broad multilingual transcription coverage +- **Speaker diarization** — Up to 32 speakers identified and labeled +- **Entity detection** — Automatic identification of names, places, and key terms +- **Keyword boost** — Improve accuracy on domain-specific vocabulary + + + + +## Available workflows + + + + Transcribe audio to text directly in your workflow. + + + Download the workflow JSON. + + diff --git a/tutorials/partner-nodes/elevenlabs/speech-to-speech.mdx b/tutorials/partner-nodes/elevenlabs/speech-to-speech.mdx new file mode 100644 index 000000000..fa42d4afc --- /dev/null +++ b/tutorials/partner-nodes/elevenlabs/speech-to-speech.mdx @@ -0,0 +1,30 @@ +--- +title: "ElevenLabs Speech to Speech" +description: "Transform an existing audio recording into a different voice or style while preserving content and delivery using ElevenLabs in ComfyUI." +sidebarTitle: "Speech to Speech" +--- + +import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; +import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; + +Take an existing audio recording and transform it into a different voice or style while preserving the spoken content and delivery. Speech to Speech retains the original intonation, pacing, and emotional delivery while applying a new voice identity. + +## Key capabilities + +- **Preserve delivery** — Retain original intonation, pacing, and emotional expression +- **Voice transformation** — Apply a different voice identity to existing audio +- **Style transfer** — Change the vocal style or character without re-recording + + + + +## Available workflows + + + + Transform existing audio into a new voice. + + + Download the workflow JSON. + + diff --git a/tutorials/partner-nodes/elevenlabs/text-to-dialogue.mdx b/tutorials/partner-nodes/elevenlabs/text-to-dialogue.mdx new file mode 100644 index 000000000..d5c721c30 --- /dev/null +++ b/tutorials/partner-nodes/elevenlabs/text-to-dialogue.mdx @@ -0,0 +1,30 @@ +--- +title: "ElevenLabs Text to Dialogue" +description: "Generate multi-character dialogue with distinct voices from a single text input using ElevenLabs in ComfyUI." +sidebarTitle: "Text to Dialogue" +--- + +import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; +import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; + +Generate multi-character dialogue with distinct voices from a single text input. Define character voices and script the conversation — ElevenLabs renders each character with its designated voice in one pass. + +## Key capabilities + +- **Multi-voice generation** — Multiple characters with distinct voices in a single pass +- **Script-driven** — Input structured dialogue text with character assignments +- **Consistent character voices** — Each character maintains its designated voice identity + + + + +## Available workflows + + + + Generate multi-voice dialogue from text. + + + Download the workflow JSON. + + diff --git a/tutorials/partner-nodes/elevenlabs/text-to-sound-effects.mdx b/tutorials/partner-nodes/elevenlabs/text-to-sound-effects.mdx new file mode 100644 index 000000000..68975f92f --- /dev/null +++ b/tutorials/partner-nodes/elevenlabs/text-to-sound-effects.mdx @@ -0,0 +1,30 @@ +--- +title: "ElevenLabs Text to Sound Effects" +description: "Generate sound effects, ambient audio, and foley from text descriptions using ElevenLabs in ComfyUI." +sidebarTitle: "Text to Sound Effects" +--- + +import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; +import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; + +Generate sound effects, ambient audio, and foley from text descriptions. Describe the audio you need — from footsteps to thunderstorms — and ElevenLabs produces the matching sound effect. + +## Key capabilities + +- **Text-prompted sound generation** — Describe the sound, get the audio +- **Ambient environments** — Generate background atmospheres and room tones +- **Foley effects** — Create custom sound effects for video and game production + + + + +## Available workflows + + + + Generate sound effects from text. + + + Download the workflow JSON. + + diff --git a/tutorials/partner-nodes/elevenlabs/voice-isolation.mdx b/tutorials/partner-nodes/elevenlabs/voice-isolation.mdx new file mode 100644 index 000000000..3343cead5 --- /dev/null +++ b/tutorials/partner-nodes/elevenlabs/voice-isolation.mdx @@ -0,0 +1,31 @@ +--- +title: "ElevenLabs Voice Isolation" +description: "Separate voice tracks from background noise, music, or ambient sound using ElevenLabs in ComfyUI." +sidebarTitle: "Voice Isolation" +--- + +import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; +import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; + +Separate the voice track from background noise, music, or ambient sound in any audio recording. ElevenLabs Voice Isolation produces clean vocal tracks from noisy source material — recordings captured on-location, in crowds, or with background music. + +## Key capabilities + +- **Clean voice extraction** — Isolate speech from any background audio +- **Noise removal** — Eliminate ambient noise, wind, and room reverb +- **Music separation** — Separate vocals from music tracks +- **High preservation** — Maintain voice quality and naturalness during isolation + + + + +## Available workflows + + + + Isolate voice from noisy audio. + + + Download the workflow JSON. + + diff --git a/tutorials/partner-nodes/grok/grok-imagine.mdx b/tutorials/partner-nodes/grok/grok-imagine.mdx new file mode 100644 index 000000000..815f9cfe2 --- /dev/null +++ b/tutorials/partner-nodes/grok/grok-imagine.mdx @@ -0,0 +1,47 @@ +--- +title: "Grok Imagine — Text to Image & Image Editing" +description: "Generate and edit images with xAI's Grok Imagine models — cinematic quality with dramatic lighting and rich color saturation." +sidebarTitle: "Grok Imagine" +--- + +import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; +import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; + +Generate images from text prompts and edit existing images using Grok Imagine by xAI. With pro/standard/beta quality tiers, Grok delivers a distinct moody, cinematic aesthetic — dramatic lighting, rich color saturation, and strong character rendering with particular strengths in anime and portraiture. + +## Key capabilities + +- **Text to Image** — Generate images from text prompts with three quality tiers +- **Image Editing** — Modify existing images using natural language instructions +- **Cinematic aesthetic** — Dramatic lighting, rich color saturation, moody atmosphere + + + + +## Available workflows + +### Text to Image + +Generate images from text prompts using the highest quality model for pro-tier results. + + + + Generate images from text instantly. + + + Download the workflow JSON. + + + +### Image Editing + +Edit existing images using text-based instructions for modifications. + + + + Edit images with text instructions. + + + Download the workflow JSON. + + diff --git a/tutorials/partner-nodes/grok/grok.mdx b/tutorials/partner-nodes/grok/grok-video.mdx similarity index 53% rename from tutorials/partner-nodes/grok/grok.mdx rename to tutorials/partner-nodes/grok/grok-video.mdx index 491a52381..ae36292a6 100644 --- a/tutorials/partner-nodes/grok/grok.mdx +++ b/tutorials/partner-nodes/grok/grok-video.mdx @@ -1,54 +1,26 @@ --- -title: "Grok Image & Video Generation" -description: "Generate images and videos with xAI's Grok models — fast, moody aesthetics with strong anime and cinematic character rendering." -sidebarTitle: "Grok" +title: "Grok Video — Text to Video, Video Editing, Extending & Reference to Video" +description: "Generate, edit, extend, and create video from reference images using xAI's Grok Video models in ComfyUI." +sidebarTitle: "Grok Video" --- import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; -Grok Imagine by xAI brings a distinct, moody visual style to AI image and video generation. It produces cinematic-quality images with dramatic lighting and rich color saturation, and supports text-to-video, video editing, and reference-to-video generation with consistent character identity. +Generate video from text descriptions, edit existing video footage, extend clips naturally, and create consistent video using reference images — all with Grok's signature moody visual style. Grok Video brings xAI's cinematic aesthetic to motion content. ## Key capabilities -- **Text to Image** — Generate images from text prompts with pro/standard/beta quality tiers -- **Image Editing** — Edit existing images with text instructions -- **Text to Video** — Generate video from text descriptions -- **Video Editing** — Edit existing video footage -- **Video Extending** — Extend an input video (15s in, extend up to 10s) -- **Reference to Video** — Generate consistent video using up to 7 reference images +- **Text to Video** — Generate video from text descriptions with cinematic quality +- **Video Editing** — Edit existing video footage with text instructions +- **Video Extending** — Extend an input video by up to 10 seconds with natural continuations +- **Reference to Video** — Generate video guided by up to 7 reference images for identity consistency ## Available workflows -### Text to Image - -Generate images from text prompts using the highest quality model for pro-tier results. - - - - Generate images from text instantly. - - - Download the workflow JSON. - - - -### Image Editing - -Edit existing images using text-based instructions for modifications. - - - - Edit images with text instructions. - - - Download the workflow JSON. - - - ### Text to Video Generate video from text descriptions with Grok's signature visual style. diff --git a/tutorials/partner-nodes/grok/overview.mdx b/tutorials/partner-nodes/grok/overview.mdx new file mode 100644 index 000000000..b18738e3e --- /dev/null +++ b/tutorials/partner-nodes/grok/overview.mdx @@ -0,0 +1,35 @@ +--- +title: "Grok Image & Video Generation" +description: "Generate images and videos with xAI's Grok models — fast, moody aesthetics with strong anime and cinematic character rendering." +sidebarTitle: "Grok" +--- + +import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; +import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; + +Grok Imagine by xAI brings a distinct, moody visual style to AI image and video generation. It produces cinematic-quality images with dramatic lighting and rich color saturation, and supports text-to-video, video editing, and reference-to-video generation with consistent character identity. + +## Key capabilities + +- **Text to Image** — Generate images from text prompts with pro/standard/beta quality tiers +- **Image Editing** — Edit existing images with text instructions +- **Text to Video** — Generate video from text descriptions +- **Video Editing** — Edit existing video footage +- **Video Extending** — Extend an input video (15s in, extend up to 10s) +- **Reference to Video** — Generate consistent video using up to 7 reference images + + + + +## Models + +Grok's creative tools are organized into two primary model groups. **Grok Imagine** handles all image generation and editing tasks, while **Grok Video** covers video generation, editing, extension, and reference-based video creation. + + + + Image generation and editing with text prompts. + + + Video generation, editing, and extension with reference support. + + diff --git a/tutorials/partner-nodes/hitpaw/hitpaw-image-enhance.mdx b/tutorials/partner-nodes/hitpaw/hitpaw-image-enhance.mdx new file mode 100644 index 000000000..d16f61299 --- /dev/null +++ b/tutorials/partner-nodes/hitpaw/hitpaw-image-enhance.mdx @@ -0,0 +1,27 @@ +--- +title: "HitPaw Image Enhance" +description: "Upscale and restore images using HitPaw's general AI enhancement model in ComfyUI." +sidebarTitle: "HitPaw Image Enhance" +--- + +import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; +import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; + +HitPaw Image Enhance is a multi-purpose AI upscaling and restoration model that handles noise, blur, artifacts, and low resolution across photographs, renders, and AI-generated images. It uses content-aware processing to automatically apply optimized enhancement strategies for natural results. + +## Key capabilities + +- **Multi-purpose upscaling** — Handles noise, blur, artifacts, and low resolution in a single pass. +- **Content-aware processing** — Automatically detects image content type (portrait, landscape, text, etc.) and applies optimized strategies. +- **No local GPU required** — All processing runs on ComfyCloud's infrastructure. +- **Minimal configuration** — Simple drop-in node with reliable results and no extensive parameter tuning. + + + + +## Available workflows + + + Launch the image enhancement workflow on ComfyCloud + Download the workflow JSON for local ComfyUI + diff --git a/tutorials/partner-nodes/hitpaw/hitpaw-video-enhance.mdx b/tutorials/partner-nodes/hitpaw/hitpaw-video-enhance.mdx new file mode 100644 index 000000000..2b0bf05fe --- /dev/null +++ b/tutorials/partner-nodes/hitpaw/hitpaw-video-enhance.mdx @@ -0,0 +1,27 @@ +--- +title: "HitPaw Video Enhance" +description: "Enhance and upscale video footage using HitPaw's AI video enhancement model in ComfyUI." +sidebarTitle: "HitPaw Video Enhance" +--- + +import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; +import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; + +HitPaw Video Enhance applies frame-by-frame AI processing to upscale and denoise video footage. It improves resolution, reduces noise, and maintains temporal smoothness across frames — suitable for compressed footage, screen recordings, and archival video. + +## Key capabilities + +- **Frame-by-frame upscaling** — AI-powered video upscaling with temporal consistency. +- **Denoising & restoration** — Removes noise and artifacts from compressed or archival footage. +- **Temporal smoothness** — Maintains consistent quality across frames without flickering or artifacts. +- **No local GPU required** — All processing runs on ComfyCloud's infrastructure. + + + + +## Available workflows + + + Launch the video enhancement workflow on ComfyCloud + Download the workflow JSON for local ComfyUI + diff --git a/tutorials/partner-nodes/hitpaw/hitpaw.mdx b/tutorials/partner-nodes/hitpaw/overview.mdx similarity index 51% rename from tutorials/partner-nodes/hitpaw/hitpaw.mdx rename to tutorials/partner-nodes/hitpaw/overview.mdx index 23c5e20ba..789db588d 100644 --- a/tutorials/partner-nodes/hitpaw/hitpaw.mdx +++ b/tutorials/partner-nodes/hitpaw/overview.mdx @@ -7,7 +7,7 @@ sidebarTitle: "HitPaw" import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; -HitPaw brings production-grade AI image and video enhancement to ComfyUI as native nodes. Its **general image enhancement** and **video enhancement** models handle upscaling, denoising, and restoration with minimal configuration, all powered by ComfyCloud with no local GPU needed. +HitPaw brings production-grade AI image and video enhancement to ComfyUI as native nodes. Its **image enhancement** and **video enhancement** models handle upscaling, denoising, and restoration with minimal configuration, all powered by ComfyCloud with no local GPU needed. ## Key capabilities @@ -21,22 +21,15 @@ HitPaw brings production-grade AI image and video enhancement to ComfyUI as nati -## Available workflows +## Models -### General Image Enhancement - -Upscale and restore images with HitPaw's all-purpose AI enhancement model. Handles low resolution, noise, blur, and compression artifacts in a single pass. - - - Launch the image enhancement workflow on ComfyCloud - Download the workflow JSON for local ComfyUI - - -### Video Enhancement - -Enhance and upscale video footage frame-by-frame using HitPaw's video AI. Improves resolution, reduces noise, and maintains temporal smoothness across frames. +HitPaw offers two AI enhancement workflows in ComfyUI — one for images and one for video. Both are designed as simple drop-in nodes with minimal configuration, suitable for integration into any pipeline. - Launch the video enhancement workflow on ComfyCloud - Download the workflow JSON for local ComfyUI + + Multi-purpose AI image upscaling and restoration. + + + Frame-by-frame video upscaling and denoising. + diff --git a/tutorials/partner-nodes/magnific/magnific.mdx b/tutorials/partner-nodes/magnific/magnific-creative-tools.mdx similarity index 52% rename from tutorials/partner-nodes/magnific/magnific.mdx rename to tutorials/partner-nodes/magnific/magnific-creative-tools.mdx index 0b72372a7..df495ce46 100644 --- a/tutorials/partner-nodes/magnific/magnific.mdx +++ b/tutorials/partner-nodes/magnific/magnific-creative-tools.mdx @@ -1,53 +1,25 @@ --- -title: "Magnific AI API Node Image Enhancement ComfyUI Official Examples" -description: "This guide covers how to use the Magnific partner nodes in ComfyUI for AI-powered image enhancement, upscaling, style transfer, relighting, and skin enhancement." -sidebarTitle: "Magnific AI" +title: "Magnific Creative Tools | ComfyUI Partner Nodes" +description: "Learn how to use Magnific Creative Tools in ComfyUI for Style Transfer, Relighting, and Skin Enhancement." +sidebarTitle: "Magnific Creative Tools" --- import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; -Magnific AI is a professional-grade image enhancement platform, now natively integrated into ComfyUI. It offers **Precise** and **Creative** upscaling (up to 16x), style transfer, relighting, and skin enhancement — bringing studio-quality post-processing directly into your workflow. +Magnific Creative Tools bring studio-quality post-processing to ComfyUI, including **Style Transfer** using reference images, **Relighting** with custom direction and color temperature, and **Skin Enhancement** for intelligent portrait retouching. ## Key capabilities -- **Precise Upscale** — High-fidelity image enlargement that preserves original details and introduces minimal new content; ideal for archival or product images. -- **Creative Upscale (16x)** — AI-powered upscaling that interprets and adds realistic detail at up to 16x magnification; the Creativity slider controls how much new detail is generated. - **Style Transfer** — Transfers the artistic style, color palette, and lighting from a reference image onto your source image for creative transformations. - **Relight** — Adjusts lighting direction, intensity, and color temperature on existing images, enabling seamless compositing into new environments. - **Skin Enhancer** — Intelligent portrait retouching that smooths skin, reduces blemishes, and evens skin tone while preserving natural facial features. ## Available workflows -### Image Upscale — Precise - -Generate a high-fidelity upscaled version of your input image with precise detail preservation. This mode adds minimal new content and stays as close to the original as possible. - - - - Run the precise image upscale workflow instantly on Comfy Cloud. - - - Download the workflow JSON file for local use. - - - -### Image Upscale — Creative - -Generate a creatively upscaled version of your input image at up to 16x magnification. The Creativity slider allows you to control how much new AI-generated detail is added — lower values stay closer to the original, while higher values produce richer, more detailed results. - - - - Run the creative image upscale workflow instantly on Comfy Cloud. - - - Download the workflow JSON file for local use. - - - ### Style Transfer Apply the visual style, color palette, and lighting mood from a reference image to your source image. This enables artistic reinterpretations — turn a photograph into an oil painting, or match the aesthetic of a cinematic frame. diff --git a/tutorials/partner-nodes/magnific/magnific-upscale.mdx b/tutorials/partner-nodes/magnific/magnific-upscale.mdx new file mode 100644 index 000000000..d66bff005 --- /dev/null +++ b/tutorials/partner-nodes/magnific/magnific-upscale.mdx @@ -0,0 +1,46 @@ +--- +title: "Magnific Upscale | ComfyUI Partner Nodes" +description: "Learn how to use Magnific Upscale in ComfyUI for Precise and Creative image upscaling up to 16x magnification." +sidebarTitle: "Magnific Upscale" +--- + +import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; +import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; + +Magnific Upscale offers two modes for enlarging your images: **Precise** for high-fidelity detail preservation, and **Creative** for AI-generated detail at up to 16x magnification. + + + + +## Key capabilities + +- **Precise Upscale** — High-fidelity image enlargement that preserves original details and introduces minimal new content; ideal for archival or product images. +- **Creative Upscale (16x)** — AI-powered upscaling that interprets and adds realistic detail at up to 16x magnification; the Creativity slider controls how much new detail is generated. + +## Available workflows + +### Image Upscale — Precise + +Generate a high-fidelity upscaled version of your input image with precise detail preservation. This mode adds minimal new content and stays as close to the original as possible. + + + + Run the precise image upscale workflow instantly on Comfy Cloud. + + + Download the workflow JSON file for local use. + + + +### Image Upscale — Creative + +Generate a creatively upscaled version of your input image at up to 16x magnification. The Creativity slider allows you to control how much new AI-generated detail is added — lower values stay closer to the original, while higher values produce richer, more detailed results. + + + + Run the creative image upscale workflow instantly on Comfy Cloud. + + + Download the workflow JSON file for local use. + + diff --git a/tutorials/partner-nodes/magnific/overview.mdx b/tutorials/partner-nodes/magnific/overview.mdx new file mode 100644 index 000000000..713971b4f --- /dev/null +++ b/tutorials/partner-nodes/magnific/overview.mdx @@ -0,0 +1,32 @@ +--- +title: "Magnific AI: Image Enhancement & Creative Tools | ComfyUI Partner Nodes" +description: "This guide covers how to use the Magnific partner nodes in ComfyUI for AI-powered image enhancement, upscaling, style transfer, relighting, and skin enhancement." +sidebarTitle: "Magnific AI" +--- + +import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; +import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; + +Magnific AI is a professional-grade image enhancement platform, now natively integrated into ComfyUI. It offers **Precise** and **Creative** upscaling (up to 16x), style transfer, relighting, and skin enhancement — bringing studio-quality post-processing directly into your workflow. + + + + +## Key capabilities + +- **Precise Upscale** — High-fidelity image enlargement that preserves original details and introduces minimal new content; ideal for archival or product images. +- **Creative Upscale (16x)** — AI-powered upscaling that interprets and adds realistic detail at up to 16x magnification; the Creativity slider controls how much new detail is generated. +- **Style Transfer** — Transfers the artistic style, color palette, and lighting from a reference image onto your source image for creative transformations. +- **Relight** — Adjusts lighting direction, intensity, and color temperature on existing images, enabling seamless compositing into new environments. +- **Skin Enhancer** — Intelligent portrait retouching that smooths skin, reduces blemishes, and evens skin tone while preserving natural facial features. + +## Models + + + + Upscale your images with Precise (detail-preserving) or Creative (AI-generated detail at up to 16x). + + + Apply Style Transfer, Relighting, and Skin Enhancement for studio-quality creative edits. + + diff --git a/tutorials/partner-nodes/minimax/minimax-01.mdx b/tutorials/partner-nodes/minimax/minimax-01.mdx new file mode 100644 index 000000000..c11f3c5d1 --- /dev/null +++ b/tutorials/partner-nodes/minimax/minimax-01.mdx @@ -0,0 +1,48 @@ +--- +title: "MiniMax-01 - Text and image to video" +description: "Generate cinematic video from text or images using MiniMax-01 with Director mode and Live motion in ComfyUI" +sidebarTitle: "MiniMax-01" +--- + +import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; +import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; + +MiniMax-01 is the first-generation MiniMax architecture, encompassing T2V-01 (text-to-video) and I2V-01 (image-to-video) models. T2V-01 offers full Director mode with 15 cinematic camera directions, while I2V-01 provides two modes — Director for camera-controlled image animation and Live for natural, organic motion that brings portraits, landscapes, and product shots to life. + +## Key capabilities + +- **T2V-01 Director mode** — 15 camera movement types including pan, tilt, dolly, crane, tracking, aerial, handheld, roll, zoom-in, zoom-out, push-in, pull-out, orbit, first-person, and static +- **I2V-01-Director** — Apply cinematic camera controls to image-initiated videos +- **I2V-01-Live** — Natural, lively motion with organic movement dynamics from a single image +- **High motion quality** — Industry-leading realism in character movement, physics, and fluid dynamics + + + + +## Available workflows + +### Text to video (T2V-01) + +Generate a video from a text prompt with full Director mode support. Choose from 15 camera movement types to craft the exact cinematic feel you need. + + + + Try the T2V-01 workflow instantly on Comfy Cloud. + + + Download the workflow JSON. + + + +### Image to video (I2V-01) + +Animate a starting image into a video clip. Two modes are available: I2V-01-Director brings camera direction (pan, zoom, tilt, etc.) to image-based generation, while I2V-01-Live prioritizes natural, organic motion. + + + + Try the I2V-01 workflow instantly on Comfy Cloud. + + + Download the workflow JSON. + + diff --git a/tutorials/partner-nodes/minimax/minimax-s2v-01.mdx b/tutorials/partner-nodes/minimax/minimax-s2v-01.mdx new file mode 100644 index 000000000..ab1411295 --- /dev/null +++ b/tutorials/partner-nodes/minimax/minimax-s2v-01.mdx @@ -0,0 +1,31 @@ +--- +title: "MiniMax S2V-01 - Subject reference to video" +description: "Generate video while maintaining character, face, or object identity using MiniMax S2V-01 subject reference in ComfyUI" +sidebarTitle: "MiniMax S2V-01" +--- + +import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; +import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; + +MiniMax S2V-01 enables subject reference-to-video generation, maintaining the identity of a specific character, face, or object throughout the clip. Supply a reference image alongside your text prompt, and S2V-01 ensures consistent appearance across the entire generated sequence — ideal for character-driven narratives, branded content, and product showcases where visual identity must remain stable from frame to frame. + +## Key capabilities + +- **Subject identity preservation** — Maintain character, face, or object identity across the entire generated video +- **Single reference image** — One input image is sufficient to guide consistent appearance +- **Text-guided generation** — Combine subject reference with text prompts for full creative control +- **Perfect for serial content** — Ideal for character-driven narratives, branded content, and product showcases + + + + +## Available workflows + + + + Try the S2V-01 workflow instantly on Comfy Cloud. + + + Download the workflow JSON. + + diff --git a/tutorials/partner-nodes/minimax/minimax.mdx b/tutorials/partner-nodes/minimax/minimax.mdx deleted file mode 100644 index dcfebefcc..000000000 --- a/tutorials/partner-nodes/minimax/minimax.mdx +++ /dev/null @@ -1,63 +0,0 @@ ---- -title: "MiniMax (Hailuo) - AI video generation" -description: "Generate high-quality video from text, images, and subject references using MiniMax's T2V-01, I2V-01, and S2V-01 models in ComfyUI" -sidebarTitle: "MiniMax (Hailuo)" ---- - -import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; -import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; - -MiniMax (Hailuo) is a leading AI video generation platform renowned for exceptional motion quality and cinematic output, now inside ComfyUI. Its **T2V-01** (text-to-video with 15 camera directions), **I2V-01** (image-to-video with live motion), and **S2V-01** (subject reference) models cover everything from film-grade narrative to character-driven content. - -## Key capabilities - -- **T2V-01 with Director mode** — 15 cinematic camera movement types (pan, tilt, dolly, crane, tracking, aerial, handheld, roll, zoom-in, zoom-out, push-in, pull-out, orbit, first-person, static) for precise shot-level direction -- **I2V-01-Director** — Apply cinematic camera controls to image-initiated videos, combining the visual foundation of a starting image with the expressive language of Director mode -- **I2V-01-Live** — Generate natural, lively motion from a single image with organic movement dynamics — ideal for bringing portraits, landscapes, and product shots to life -- **S2V-01 subject reference** — Maintain character, face, or object identity across the generated video using a single reference image; perfect for serial content -- **High motion quality** — Industry-leading realism in character movement, physics, and fluid dynamics across all model variants -- **Flexible input modes** — Text-only, image+text, and subject reference+text, each with optional camera direction - - - - -## Available workflows - -### Text to video (T2V-01) - -Generate a video from a text prompt with full Director mode support. Choose from 15 camera movement types — pan, tilt, dolly, crane, tracking, aerial, handheld, roll, zoom-in, zoom-out, push-in, pull-out, orbit, first-person, and static — to craft the exact cinematic feel you need. - - - - Try the T2V-01 workflow instantly on Comfy Cloud. - - - Download the workflow JSON. - - - -### Image to video (I2V-01) - -Animate a starting image into a video clip. Two modes are available: I2V-01-Director brings camera direction (pan, zoom, tilt, etc.) to image-based generation, while I2V-01-Live prioritizes natural, organic motion — perfect for bringing portraits, landscapes, and product shots to life with subtle movement and atmospheric breathing. - - - - Try the I2V-01 workflow instantly on Comfy Cloud. - - - Download the workflow JSON. - - - -### Subject reference to video (S2V-01) - -Generate a video while maintaining the identity of a specific character, face, or object throughout the clip. Supply a reference image alongside your text prompt, and S2V-01 ensures consistent appearance across the entire generated sequence. This is especially useful for character-driven narratives, branded content, and product showcases where visual identity must remain stable from frame to frame. - - - - Try the S2V-01 workflow instantly on Comfy Cloud. - - - Download the workflow JSON. - - diff --git a/tutorials/partner-nodes/minimax/overview.mdx b/tutorials/partner-nodes/minimax/overview.mdx new file mode 100644 index 000000000..50b33cd22 --- /dev/null +++ b/tutorials/partner-nodes/minimax/overview.mdx @@ -0,0 +1,32 @@ +--- +title: "MiniMax (Hailuo) - AI video generation" +description: "Generate high-quality video from text, images, and subject references using MiniMax's T2V-01, I2V-01, and S2V-01 models in ComfyUI" +sidebarTitle: "MiniMax (Hailuo)" +--- + +import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; +import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; + +MiniMax (Hailuo) is a leading AI video generation platform renowned for exceptional motion quality and cinematic output, now inside ComfyUI. It offers two model lines — **MiniMax-01** (T2V-01 text-to-video and I2V-01 image-to-video with Director/Live modes) and **MiniMax S2V-01** (subject reference-to-video) — covering everything from film-grade narrative to character-driven content. + +## Key capabilities + +- **T2V-01 with Director mode** — 15 cinematic camera movement types (pan, tilt, dolly, crane, tracking, aerial, handheld, roll, zoom-in, zoom-out, push-in, pull-out, orbit, first-person, static) for precise shot-level direction +- **I2V-01-Director** — Apply cinematic camera controls to image-initiated videos, combining the visual foundation of a starting image with the expressive language of Director mode +- **I2V-01-Live** — Generate natural, lively motion from a single image with organic movement dynamics — ideal for bringing portraits, landscapes, and product shots to life +- **S2V-01 subject reference** — Maintain character, face, or object identity across the generated video using a single reference image; perfect for serial content +- **High motion quality** — Industry-leading realism in character movement, physics, and fluid dynamics across all model variants + + + + +## Models + + + + T2V-01 with Director mode and I2V-01 with Director/Live modes — the first-generation MiniMax architecture for text and image-based video generation. + + + Subject reference-to-video for maintaining character, face, or object identity across generated clips. + + diff --git a/tutorials/partner-nodes/pixverse/overview.mdx b/tutorials/partner-nodes/pixverse/overview.mdx new file mode 100644 index 000000000..42cbef9cc --- /dev/null +++ b/tutorials/partner-nodes/pixverse/overview.mdx @@ -0,0 +1,27 @@ +--- +title: "PixVerse AI: Video Generation | ComfyUI Partner Nodes" +description: "This guide covers how to use the PixVerse partner nodes in ComfyUI for AI-powered video generation from text prompts, images, and template-based workflows." +sidebarTitle: "PixVerse AI" +--- + +import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; +import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; + +PixVerse is an AI video generation platform that produces cinematic videos from text and images, now integrated as partner nodes in ComfyUI. With support for **V5.6**, **V6**, **C1**, and **R1** models, it enables **Text-to-Video**, **Image-to-Video**, and **Template Image-to-Video** modes for rapid, repeatable video creation without separate API keys or external setup. + + + + +## Key capabilities + +- **Text-to-Video** — Generate original videos from text prompts, ideal for concept visualization, storyboarding, and rapid creative exploration. +- **Image-to-Video** — Animate a reference image while preserving subject identity and scene composition, bringing static visuals to life with natural motion. +- **Template Image-to-Video** — Apply pre-defined motion templates to input images for predictable, repeatable animation styles suitable for production pipelines. + +## Models + + + + Generate videos from text prompts, images, or templates using PixVerse's AI models. + + diff --git a/tutorials/partner-nodes/pixverse/pixverse.mdx b/tutorials/partner-nodes/pixverse/pixverse-video.mdx similarity index 81% rename from tutorials/partner-nodes/pixverse/pixverse.mdx rename to tutorials/partner-nodes/pixverse/pixverse-video.mdx index 30100f74d..9f4e93d89 100644 --- a/tutorials/partner-nodes/pixverse/pixverse.mdx +++ b/tutorials/partner-nodes/pixverse/pixverse-video.mdx @@ -1,13 +1,13 @@ --- -title: "PixVerse AI API Node Video Generation ComfyUI Official Examples" -description: "This guide covers how to use the PixVerse partner nodes in ComfyUI for AI-powered video generation from text prompts, images, and template-based workflows." -sidebarTitle: "PixVerse AI" +title: "PixVerse Video | ComfyUI Partner Nodes" +description: "Learn how to use PixVerse Video in ComfyUI for Text-to-Video, Image-to-Video, and Template Image-to-Video generation." +sidebarTitle: "PixVerse Video" --- import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; -PixVerse is an AI video generation platform that produces cinematic videos from text and images, now integrated as partner nodes in ComfyUI. It supports **Text-to-Video**, **Image-to-Video**, and **Template Image-to-Video** modes for rapid, repeatable video creation without separate API keys or external setup. +PixVerse Video enables cinematic video creation directly in ComfyUI through **Text-to-Video**, **Image-to-Video**, and **Template Image-to-Video** modes, powered by models including V5.6, V6, C1, and R1. diff --git a/tutorials/partner-nodes/quiver/overview.mdx b/tutorials/partner-nodes/quiver/overview.mdx new file mode 100644 index 000000000..ac7bf08fc --- /dev/null +++ b/tutorials/partner-nodes/quiver/overview.mdx @@ -0,0 +1,29 @@ +--- +title: "Quiver AI: SVG Generation | ComfyUI Partner Nodes" +description: "This guide covers how to use the Quiver partner nodes in ComfyUI for AI-powered SVG vector graphics generation from text prompts and reference images." +sidebarTitle: "Quiver AI" +--- + +import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; +import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; + +Quiver is an AI-powered vector graphics platform that generates high-quality SVGs from text descriptions and raster images, now available as partner nodes in ComfyUI. With models from **Arrow 1.1** (fast) to **Arrow 1.1 Max** (high-fidelity) and **Arrow Preview** (experimental), it produces resolution-independent vector artwork without manual editing. + + + + +## Key capabilities + +- **Text-to-SVG** — Generate fully editable SVG vector graphics from natural language descriptions; perfect for icons, logos, illustrations, and UI elements. +- **Image-to-SVG** — Convert raster images (PNG, JPG) into clean, scalable SVG vector graphics by tracing shapes, colors, and gradients. +- **Arrow 1.1** — Standard model offering fast and efficient SVG generation suitable for most use cases. +- **Arrow 1.1 Max** — High-quality variant for detailed and accurate vector output on complex illustrations. +- **Arrow Preview** — Early access to experimental features and upcoming model improvements. + +## Models + + + + Generate editable SVG vector graphics from text descriptions or convert raster images into scalable vector art. + + diff --git a/tutorials/partner-nodes/quiver/quiver.mdx b/tutorials/partner-nodes/quiver/quiver-svg.mdx similarity index 66% rename from tutorials/partner-nodes/quiver/quiver.mdx rename to tutorials/partner-nodes/quiver/quiver-svg.mdx index fb348d10d..e0cac18c2 100644 --- a/tutorials/partner-nodes/quiver/quiver.mdx +++ b/tutorials/partner-nodes/quiver/quiver-svg.mdx @@ -1,13 +1,13 @@ --- -title: "Quiver AI API Node SVG Generation ComfyUI Official Examples" -description: "This guide covers how to use the Quiver partner nodes in ComfyUI for AI-powered SVG vector graphics generation from text prompts and reference images." -sidebarTitle: "Quiver AI" +title: "Quiver SVG | ComfyUI Partner Nodes" +description: "Learn how to use Quiver SVG in ComfyUI for Text-to-SVG and Image-to-SVG vector graphics generation." +sidebarTitle: "Quiver SVG" --- import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; -Quiver is an AI-powered vector graphics platform that generates high-quality SVGs from text descriptions and raster images, now available as partner nodes in ComfyUI. With models from **Arrow 1.1** (fast) to **Arrow 1.1 Max** (high-fidelity), it produces resolution-independent vector artwork without manual editing. +Quiver SVG generates resolution-independent vector graphics directly in ComfyUI through **Text-to-SVG** and **Image-to-SVG** modes. With models including Arrow 1.1, Arrow 1.1 Max, and Arrow Preview, it handles everything from quick icon generation to complex illustration vectorization. @@ -16,9 +16,6 @@ Quiver is an AI-powered vector graphics platform that generates high-quality SVG - **Text-to-SVG** — Generate fully editable SVG vector graphics from natural language descriptions; perfect for icons, logos, illustrations, and UI elements. - **Image-to-SVG** — Convert raster images (PNG, JPG) into clean, scalable SVG vector graphics by tracing shapes, colors, and gradients. -- **Arrow 1.1** — Standard model offering fast and efficient SVG generation suitable for most use cases. -- **Arrow 1.1 Max** — High-quality variant for detailed and accurate vector output on complex illustrations. -- **Arrow Preview** — Early access to experimental features and upcoming model improvements. ## Available workflows diff --git a/tutorials/partner-nodes/topaz/overview.mdx b/tutorials/partner-nodes/topaz/overview.mdx new file mode 100644 index 000000000..765d3c8f5 --- /dev/null +++ b/tutorials/partner-nodes/topaz/overview.mdx @@ -0,0 +1,36 @@ +--- +title: "Image & Video Enhancement with Topaz Labs" +description: "Upscale, enhance, and restore images and video using Topaz Labs AI models — including Reimagine creative upscaling, face enhancement, Starlight diffusion upscale, and Apollo frame interpolation — all within ComfyUI." +sidebarTitle: "Topaz Labs" +--- + +import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; +import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; + +Topaz Labs brings enterprise-grade AI image and video enhancement to ComfyUI — trusted by professionals worldwide. Its suite covers **image enhancement**, **landscape upscaling**, **illustration upscaling**, and **video enhancement**, all powered by ComfyCloud with no local GPU needed. + +## Key capabilities + +- **Reimagine creative upscaling** — Guide the upscaling process with text prompts to creatively reinterpret and enhance images, not just enlarge them. +- **Face enhancement** — Dedicated facial detail reconstruction model that detects and sharpens faces in images and video, preserving natural skin texture and expression. +- **8K output support** — Upscale images and video frames to ultra-high 8K resolution while maintaining sharpness and visual coherence. +- **Starlight diffusion upscaler** — Diffusion-based upscaling optimized for landscape photography, hallucinating realistic detail like foliage, rock textures, and clouds. +- **Apollo frame interpolation** — Generate smooth intermediate frames between existing video frames for slow-motion effects or frame rate conversion. +- **Illustration-optimized upscale** — Specialized model for flat-colored artwork, line art, and digital illustrations that preserves hard edges and clean gradients. +- **Video enhancement pipeline** — Comprehensive video processing including denoising, deinterlacing, stabilization, and intelligent frame-by-frame upscaling. + + + + +## Models + +Topaz Labs offers a suite of AI enhancement models for both image and video. The **Image Enhance** group covers general upscaling, landscape diffusion upscaling, and illustration-optimized upscaling, while **Video Enhance** provides frame-by-frame enhancement with Apollo interpolation. + + + + General image upscaling, Starlight landscape diffusion, and illustration upscaling. + + + Frame-by-frame video upscaling, denoising, and Apollo interpolation. + + diff --git a/tutorials/partner-nodes/topaz/topaz-image.mdx b/tutorials/partner-nodes/topaz/topaz-image.mdx new file mode 100644 index 000000000..c90f1d77d --- /dev/null +++ b/tutorials/partner-nodes/topaz/topaz-image.mdx @@ -0,0 +1,50 @@ +--- +title: "Topaz Image Enhance" +description: "Upscale, enhance, and restore images using Topaz Labs AI — including general image enhancement, Starlight landscape diffusion upscaling, and illustration-optimized upscaling in ComfyUI." +sidebarTitle: "Topaz Image Enhance" +--- + +import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; +import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; + +Topaz Image Enhance offers three specialized workflows for AI-powered image upscaling and restoration. The **general enhancement** model handles photographs and renders, the **Starlight landscape upscaler** uses diffusion for photorealistic landscape detail, and the **illustration upscale** preserves hard edges and clean gradients for digital art. + +## Key capabilities + +- **General-purpose enhancement** — AI upscaling and denoising suitable for photographs, renders, and most image types. +- **Starlight landscape diffusion** — Diffusion-based upscaling that hallucinates realistic natural detail like foliage, rock textures, and clouds at up to 8K resolution. +- **Illustration-optimized upscaling** — Specialized model for flat-colored artwork, line art, and digital illustrations that preserves hard edges and clean color gradients. +- **Face enhancement** — Dedicated facial detail reconstruction that preserves natural skin texture and expression. +- **8K output support** — Upscale images to ultra-high 8K resolution while maintaining sharpness and visual coherence. + + + + +## Available workflows + +### Image Enhancement + +General-purpose AI image upscaling and denoising using Topaz's core enhancement model. Suitable for photographs, renders, and most image types. + + + Launch the image enhancement workflow on ComfyCloud + Download the workflow JSON for local ComfyUI + + +### Landscape Upscaler (Starlight) + +Diffusion-based upscaling specialized for landscape photography. Uses the Starlight model to hallucinate realistic natural detail at up to 8K resolution. + + + Launch the Starlight landscape upscaler on ComfyCloud + Download the workflow JSON for local ComfyUI + + +### Illustration Upscale + +Optimized upscaling for digital illustrations, anime art, and flat-color artwork. Preserves hard edges, crisp lines, and smooth color gradients. + + + Launch the illustration upscale workflow on ComfyCloud + Download the workflow JSON for local ComfyUI + diff --git a/tutorials/partner-nodes/topaz/topaz-video-enhance.mdx b/tutorials/partner-nodes/topaz/topaz-video-enhance.mdx new file mode 100644 index 000000000..140a3bfba --- /dev/null +++ b/tutorials/partner-nodes/topaz/topaz-video-enhance.mdx @@ -0,0 +1,28 @@ +--- +title: "Topaz Video Enhance" +description: "Enhance, upscale, and interpolate video using Topaz Labs AI — including frame-by-frame upscaling, denoising, deinterlacing, stabilization, and Apollo frame interpolation in ComfyUI." +sidebarTitle: "Topaz Video Enhance" +--- + +import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; +import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; + +Topaz Video Enhance applies AI-powered frame-by-frame processing to upscale, denoise, and restore video footage. It supports deinterlacing, stabilization, and the Apollo frame interpolation model for creating smooth slow-motion effects. + +## Key capabilities + +- **Frame-by-frame upscaling** — Intelligent AI upscaling that maintains consistency across video frames. +- **Denoising & restoration** — Removes noise, artifacts, and compression artifacts from video footage. +- **Deinterlacing & stabilization** — Converts interlaced video and stabilizes shaky footage. +- **Apollo frame interpolation** — Generates smooth intermediate frames for slow-motion effects or frame rate conversion. +- **8K output support** — Upscale video frames to ultra-high 8K resolution. + + + + +## Available workflows + + + Launch the video enhancement workflow on ComfyCloud + Download the workflow JSON for local ComfyUI + diff --git a/tutorials/partner-nodes/topaz/topaz.mdx b/tutorials/partner-nodes/topaz/topaz.mdx deleted file mode 100644 index f9b21f2d7..000000000 --- a/tutorials/partner-nodes/topaz/topaz.mdx +++ /dev/null @@ -1,61 +0,0 @@ ---- -title: "Image & Video Enhancement with Topaz Labs" -description: "Upscale, enhance, and restore images and video using Topaz Labs AI models — including Reimagine creative upscaling, face enhancement, Starlight diffusion upscale, and Apollo frame interpolation — all within ComfyUI." -sidebarTitle: "Topaz Labs" ---- - -import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; -import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; - -Topaz Labs brings enterprise-grade AI image and video enhancement to ComfyUI — trusted by professionals worldwide. Its suite covers **upscaling**, **denoising**, **face enhancement**, **Starlight diffusion upscale** for landscapes, **Reimagine** creative upscaling, and the **Apollo** frame interpolation model for video. - -## Key capabilities - -- **Reimagine creative upscaling** — Guide the upscaling process with text prompts to creatively reinterpret and enhance images, not just enlarge them. -- **Face enhancement** — Dedicated facial detail reconstruction model that detects and sharpens faces in images and video, preserving natural skin texture and expression. -- **8K output support** — Upscale images and video frames to ultra-high 8K resolution while maintaining sharpness and visual coherence. -- **Starlight diffusion upscaler** — Diffusion-based upscaling optimized for landscape photography, hallucinating realistic detail like foliage, rock textures, and clouds. -- **Apollo frame interpolation** — Generate smooth intermediate frames between existing video frames for slow-motion effects or frame rate conversion. -- **Illustration-optimized upscale** — Specialized model for flat-colored artwork, line art, and digital illustrations that preserves hard edges and clean gradients. -- **Video enhancement pipeline** — Comprehensive video processing including denoising, deinterlacing, stabilization, and intelligent frame-by-frame upscaling. - - - - -## Available workflows - -### Image Enhancement - -General-purpose AI image upscaling and denoising using Topaz's core enhancement model. Suitable for photographs, renders, and most image types. - - - Launch the image enhancement workflow on ComfyCloud - Download the workflow JSON for local ComfyUI - - -### Video Enhancement - -Frame-by-frame AI video upscaling, denoising, and stabilization. Supports deinterlacing and Apollo frame interpolation for smooth slow-motion. - - - Launch the video enhancement workflow on ComfyCloud - Download the workflow JSON for local ComfyUI - - -### Landscape Upscaler (Starlight) - -Diffusion-based upscaling specialized for landscape photography. Uses the Starlight model to hallucinate realistic natural detail at up to 8K resolution. - - - Launch the Starlight landscape upscaler on ComfyCloud - Download the workflow JSON for local ComfyUI - - -### Illustration Upscale - -Optimized upscaling for digital illustrations, anime art, and flat-color artwork. Preserves hard edges, crisp lines, and smooth color gradients. - - - Launch the illustration upscale workflow on ComfyCloud - Download the workflow JSON for local ComfyUI - diff --git a/tutorials/partner-nodes/veo/overview.mdx b/tutorials/partner-nodes/veo/overview.mdx new file mode 100644 index 000000000..280b6a58f --- /dev/null +++ b/tutorials/partner-nodes/veo/overview.mdx @@ -0,0 +1,35 @@ +--- +title: "Video Generation with Google Veo" +description: "Generate cinematic, high-resolution videos using Google DeepMind's Veo models — Veo 3.1 and Veo 2.0 — directly within ComfyUI." +sidebarTitle: "Veo" +--- + +import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; +import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; + +Google DeepMind's **Veo 3.1** and **Veo 2.0** bring state-of-the-art AI video generation to ComfyUI. Veo 3.1 delivers cinema-grade 4K video with native audio, while Veo 2.0 offers fast image-to-video for rapid prototyping — all seamlessly integrated into your ComfyUI pipeline. + +## Key capabilities + +- **Cinema-grade 4K video** — Veo 3.1 generates 4K resolution video with crisp detail, natural lighting, and realistic physics, suitable for professional film and advertising use. +- **Native audio sync** — Video output includes synchronized audio, eliminating the need for separate audio generation and manual alignment in post-production. +- **Text-to-video (Veo 3.1)** — Generate high-quality video directly from text prompts, with support for detailed scene direction, camera movement, and visual style control. +- **Optimized speed tier** — A faster inference variant of Veo 3.1 that trades slight quality for significantly reduced generation time, ideal for iteration and draft production. +- **Image-to-video (Veo 2.0)** — Animate static images into short video clips, preserving the original composition while adding natural motion. +- **Seamless ComfyUI pipeline** — Full integration allows chaining Veo output with ComfyUI's existing upscaling, audio, and compositing nodes without file export steps. + + + + +## Models + +Google Veo offers two model generations in ComfyUI, each optimized for different use cases. Veo 3.1 is the flagship text-to-video model with two speed tiers, while Veo 2.0 provides a lightweight image-to-video option. + + + + Cinema-grade 4K text-to-video with native audio. Standard and Fast tiers available. + + + Lightweight image-to-video for animating static images. + + diff --git a/tutorials/partner-nodes/veo/veo-2-0.mdx b/tutorials/partner-nodes/veo/veo-2-0.mdx new file mode 100644 index 000000000..0f2be174c --- /dev/null +++ b/tutorials/partner-nodes/veo/veo-2-0.mdx @@ -0,0 +1,26 @@ +--- +title: "Veo 2.0 — Image-to-Video" +description: "Animate static images into short video clips using Google DeepMind's Veo 2.0 in ComfyUI." +sidebarTitle: "Veo 2.0" +--- + +import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; +import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; + +Veo 2.0 is Google DeepMind's lightweight image-to-video model in ComfyUI. It animates static images into short video clips, preserving the original composition while adding natural motion — ideal for breathing life into concept art, AI-generated images, or photographs. + +## Key capabilities + +- **Image-to-video animation** — Convert static images into short, natural-looking video clips. +- **Composition preservation** — Maintains the original image structure, colors, and layout while adding motion. +- **Natural motion generation** — Produces realistic movement that respects the source image's content and style. + + + + +## Available workflows + + + Launch the Veo 2.0 I2V workflow on ComfyCloud + Download the workflow JSON for local ComfyUI + diff --git a/tutorials/partner-nodes/veo/veo-3-1.mdx b/tutorials/partner-nodes/veo/veo-3-1.mdx new file mode 100644 index 000000000..e9517d37f --- /dev/null +++ b/tutorials/partner-nodes/veo/veo-3-1.mdx @@ -0,0 +1,40 @@ +--- +title: "Veo 3.1 — Text-to-Video" +description: "Generate cinema-grade 4K video with native audio using Google DeepMind's Veo 3.1 in ComfyUI — available in Standard and Fast tiers." +sidebarTitle: "Veo 3.1" +--- + +import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; +import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; + +Veo 3.1 is Google DeepMind's flagship text-to-video model, delivering cinema-grade 4K video with synchronized native audio. Available in **Standard** (full quality) and **Fast** (optimized speed) tiers, it supports detailed scene direction, camera movement, and visual style control. + +## Key capabilities + +- **Cinema-grade 4K video** — Generates 4K resolution video with crisp detail, natural lighting, and realistic physics. +- **Native audio sync** — Video output includes synchronized audio, eliminating the need for post-production alignment. +- **Text-to-video generation** — Create high-quality video directly from text prompts with full creative control. +- **Optimized Fast tier** — A faster inference variant that trades slight quality for significantly reduced generation time, ideal for iteration and draft production. + + + + +## Available workflows + +### Veo 3.1 — Text-to-Video (Standard) + +Generate full-quality 4K video with native audio from a text prompt. Best for final production output where quality matters most. + + + Launch the full-quality Veo 3.1 workflow on ComfyCloud + Download the workflow JSON for local ComfyUI + + +### Veo 3.1 — Text-to-Video (Fast) + +A lightweight variant of Veo 3.1 optimized for speed. Use this for rapid ideation, draft generation, and iterative prompt refinement. + + + Launch the fast Veo 3.1 workflow on ComfyCloud + Download the workflow JSON for local ComfyUI + diff --git a/tutorials/partner-nodes/veo/veo.mdx b/tutorials/partner-nodes/veo/veo.mdx deleted file mode 100644 index 53a562042..000000000 --- a/tutorials/partner-nodes/veo/veo.mdx +++ /dev/null @@ -1,51 +0,0 @@ ---- -title: "Video Generation with Google Veo" -description: "Generate cinematic, high-resolution videos using Google DeepMind's Veo models — Veo 3.1 and Veo 2.0 — directly within ComfyUI." -sidebarTitle: "Veo" ---- - -import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; -import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; - -Google DeepMind's **Veo 3.1** and **Veo 2.0** bring state-of-the-art AI video generation to ComfyUI. Veo 3.1 delivers cinema-grade 4K video with native audio, while Veo 2.0 offers fast text-to-video and image-to-video for rapid prototyping — all seamlessly integrated into your ComfyUI pipeline. - -## Key capabilities - -- **Cinema-grade 4K video** — Veo 3.1 generates 4K resolution video with crisp detail, natural lighting, and realistic physics, suitable for professional film and advertising use. -- **Native audio sync** — Video output includes synchronized audio, eliminating the need for separate audio generation and manual alignment in post-production. -- **Text-to-video (Veo 3.1 & 2.0)** — Generate high-quality video directly from text prompts, with support for detailed scene direction, camera movement, and visual style control. -- **Image-to-video (Veo 2.0)** — Animate static images into short video clips, preserving the original composition while adding natural motion. -- **Optimized speed tier (api_veo3_fast)** — A faster inference variant of Veo 3.1 that trades slight quality for significantly reduced generation time, ideal for iteration and draft production. -- **Seamless ComfyUI pipeline** — Full integration allows chaining Veo output with ComfyUI's existing upscaling, audio, and compositing nodes without file export steps. - - - - -## Available workflows - -### Veo 3.1 — Text-to-Video (Standard) - -Generate full-quality 4K video with native audio from a text prompt. Best for final production output where quality matters most. - - - Launch the full-quality Veo 3.1 workflow on ComfyCloud - Download the workflow JSON for local ComfyUI - - -### Veo 3.1 — Text-to-Video (Fast) - -A lightweight variant of Veo 3.1 optimized for speed. Use this for rapid ideation, draft generation, and iterative prompt refinement. - - - Launch the fast Veo 3.1 workflow on ComfyCloud - Download the workflow JSON for local ComfyUI - - -### Veo 2.0 — Image-to-Video - -Animate a static image into a short video clip. Great for breathing life into concept art, AI-generated images, or photographs. - - - Launch the Veo 2.0 I2V workflow on ComfyCloud - Download the workflow JSON for local ComfyUI - diff --git a/tutorials/partner-nodes/vidu/overview.mdx b/tutorials/partner-nodes/vidu/overview.mdx new file mode 100644 index 000000000..b296e788d --- /dev/null +++ b/tutorials/partner-nodes/vidu/overview.mdx @@ -0,0 +1,36 @@ +--- +title: "Vidu - AI video generation" +description: "Generate high-quality video from text, images, and reference materials using Vidu's Q1, Q2, and Q3 model generations in ComfyUI" +sidebarTitle: "Vidu" +--- + +import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; +import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; + +Vidu is a fast AI video generation platform by Shengshu Technology, now available inside ComfyUI. Its three model generations — **Q1**, **Q2**, and **Q3** — deliver a range of speed and quality options, from foundation generation with broad input support to 3x faster inference with camera control and exceptional anime output driven by the latest Vidu engines (2.0, 3.0, and 3.5). + +## Key capabilities + +- **Multi-generation evolution** — Choose between Q1 (foundation), Q2 (fast + multi-reference + camera control), and Q3 (latest narrative + anime/2D focus) to match your workflow +- **Fast inference** — Q3 generates 4-second clips in as fast as 10 seconds; Q2 delivers 3x the speed of Q1 +- **Rich input modes** — Text-to-video, image-to-video, reference-to-video (up to 7 images), start-end-to-video, video extension, and first-last-frame-to-video +- **Camera language (Q2)** — Pan, zoom, rotation, and other cinematic moves for precise shot direction +- **Exceptional anime & 2D animation (Q3)** — Industry-leading output quality for animation production pipelines +- **Subject consistency** — Maintain character and scene identity across multi-reference inputs + + + + +## Models + + + + Foundation generation with text-to-video, image-to-video, reference-to-video and more — the versatile starting point. + + + 3x faster inference with multi-reference support and cinematic camera control — ideal for iterative workflows. + + + The latest generation with improved narrative understanding, outstanding anime/2D quality, and Vidu 3.5 engine. + + diff --git a/tutorials/partner-nodes/vidu/vidu-q1.mdx b/tutorials/partner-nodes/vidu/vidu-q1.mdx new file mode 100644 index 000000000..45339f053 --- /dev/null +++ b/tutorials/partner-nodes/vidu/vidu-q1.mdx @@ -0,0 +1,77 @@ +--- +title: "Vidu Q1 - Foundation video generation" +description: "Generate video from text, images, reference images, start-end frames, and extend existing clips using Vidu Q1 in ComfyUI" +sidebarTitle: "Vidu Q1" +--- + +import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; +import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; + +Vidu Q1 is the first-generation Vidu model that handles all core video generation types. It offers a broad range of input modes — text, image, reference images, start-and-end frames, and video extension — making it a reliable choice for general-purpose production where speed and expressiveness are balanced. + +## Key capabilities + +- **Five generation modes** — Text-to-video, image-to-video, reference-to-video (up to 7 images), start-end-to-video, and video extension +- **Flexible subject reference** — Maintain character and scene identity across multi-reference inputs +- **Video extension** — Extend existing video clips with coherent continuation +- **Balanced performance** — Reliable quality suitable for a wide range of production scenarios + + + + +## Available workflows + +### Text to video + + + + Try the Q1 Text-to-Video workflow instantly on Comfy Cloud. + + + Download the workflow JSON. + + + +### Image to video + + + + Try the Q1 Image-to-Video workflow instantly on Comfy Cloud. + + + Download the workflow JSON. + + + +### Reference to video + + + + Try the Q1 Reference-to-Video workflow instantly on Comfy Cloud. + + + Download the workflow JSON. + + + +### Start end to video + + + + Try the Q1 Start-End-to-Video workflow instantly on Comfy Cloud. + + + Download the workflow JSON. + + + +### Video extension + + + + Try the Q1 Video Extension workflow instantly on Comfy Cloud. + + + Download the workflow JSON. + + diff --git a/tutorials/partner-nodes/vidu/vidu-q2.mdx b/tutorials/partner-nodes/vidu/vidu-q2.mdx new file mode 100644 index 000000000..6e5fbf46b --- /dev/null +++ b/tutorials/partner-nodes/vidu/vidu-q2.mdx @@ -0,0 +1,66 @@ +--- +title: "Vidu Q2 - Fast generation with camera control" +description: "Generate video 3x faster with multi-reference support and cinematic camera controls using Vidu Q2 in ComfyUI" +sidebarTitle: "Vidu Q2" +--- + +import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; +import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; + +Vidu Q2 builds on the Q1 foundation with significantly faster inference (3x speedup), support for up to 7 reference images, enhanced subject consistency, and cinematic camera language controls including pan, zoom, and rotation. It is available as text-to-video, image-to-video, reference-to-video, and first-last-frame-to-video, making it ideal for iterative workflows and shot-level direction. + +## Key capabilities + +- **3x faster inference** — Significantly reduced generation time compared to Q1 +- **Multi-reference support** — Up to 7 reference images for enhanced subject consistency +- **Cinematic camera control** — Pan, zoom, rotation, and other camera language for precise shot direction +- **Four generation modes** — Text-to-video, image-to-video, reference-to-video, and first-last-frame-to-video + + + + +## Available workflows + +### Text to video + + + + Try the Q2 Text-to-Video workflow instantly on Comfy Cloud. + + + Download the workflow JSON. + + + +### Image to video + + + + Try the Q2 Image-to-Video workflow instantly on Comfy Cloud. + + + Download the workflow JSON. + + + +### Reference to video + + + + Try the Q2 Reference-to-Video workflow instantly on Comfy Cloud. + + + Download the workflow JSON. + + + +### First last frame to video + + + + Try the Q2 First-Last-Frame-to-Video workflow instantly on Comfy Cloud. + + + Download the workflow JSON. + + diff --git a/tutorials/partner-nodes/vidu/vidu-q3.mdx b/tutorials/partner-nodes/vidu/vidu-q3.mdx new file mode 100644 index 000000000..54ff4ede9 --- /dev/null +++ b/tutorials/partner-nodes/vidu/vidu-q3.mdx @@ -0,0 +1,44 @@ +--- +title: "Vidu Q3 - Latest generation for narrative and animation" +description: "Generate high-quality video with improved narrative understanding and outstanding anime output using Vidu Q3 in ComfyUI" +sidebarTitle: "Vidu Q3" +--- + +import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; +import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; + +Vidu Q3 is the newest and most capable Vidu model generation, powered by the Vidu 3.5 engine. It delivers the highest visual quality with a particular strength in anime and 2D animation styles, making it the top choice for animation production pipelines. Q3 also introduces improved narrative understanding for more coherent storytelling. Generation is remarkably fast — a 4-second clip can be produced in as little as 10 seconds. + +## Key capabilities + +- **Vidu 3.5 engine** — The latest engine delivering the highest visual quality across the Vidu family +- **Exceptional anime quality** — Industry-leading output for 2D animation production +- **Improved narrative understanding** — More coherent storytelling across generated clips +- **Fast generation speed** — 4-second clips in as fast as 10 seconds + + + + +## Available workflows + +### Text to video + + + + Try the Q3 Text-to-Video workflow instantly on Comfy Cloud. + + + Download the workflow JSON. + + + +### Image to video + + + + Try the Q3 Image-to-Video workflow instantly on Comfy Cloud. + + + Download the workflow JSON. + + diff --git a/tutorials/partner-nodes/vidu/vidu.mdx b/tutorials/partner-nodes/vidu/vidu.mdx deleted file mode 100644 index 3fc55f1b2..000000000 --- a/tutorials/partner-nodes/vidu/vidu.mdx +++ /dev/null @@ -1,135 +0,0 @@ ---- -title: "Vidu - AI video generation" -description: "Generate high-quality video from text, images, and reference materials using Vidu's Q1, Q2, and Q3 model generations in ComfyUI" -sidebarTitle: "Vidu" ---- - -import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; -import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; - -Vidu is a fast AI video generation platform by Shengshu Technology, now available inside ComfyUI. Its **Q1**, **Q2**, and **Q3** model families deliver a range of speed and quality options — from foundation generation to 3x faster inference with camera control and exceptional anime output. - -## Key capabilities - -- **Multi-model generation** — Choose between Q1 (foundation), Q2 (fast + multi-reference + camera control), and Q3 (latest narrative + anime/2D focus) to match your workflow -- **Fast inference** — Q3 generates 4-second clips in as fast as 10 seconds; Q2 delivers 3x the speed of Q1 -- **Rich input modes** — Text-to-video, image-to-video (with optional prompt), reference-to-video (up to 7 images), start-end-to-video, and video extension -- **Camera language (Q2)** — Pan, zoom, rotation, and other cinematic moves for precise shot direction -- **Exceptional anime & 2D animation (Q3)** — Industry-leading output quality for animation production pipelines -- **Subject consistency** — Maintain character and scene identity across multi-reference inputs - - - - -## Available workflows - -### Q1 — Foundation generation - -The first-generation Vidu model handles all core generation types: text-to-video, image-to-video, reference-to-video, start-end-to-video, and video extension. Reliable for general-purpose production where speed and expressiveness are balanced. - - - - Try the Q1 Text-to-Video workflow instantly on Comfy Cloud. - - - Download the workflow JSON. - - - - - - Try the Q1 Image-to-Video workflow instantly on Comfy Cloud. - - - Download the workflow JSON. - - - - - - Try the Q1 Reference-to-Video workflow instantly on Comfy Cloud. - - - Download the workflow JSON. - - - - - - Try the Q1 Start-End-to-Video workflow instantly on Comfy Cloud. - - - Download the workflow JSON. - - - - - - Try the Q1 Video Extension workflow instantly on Comfy Cloud. - - - Download the workflow JSON. - - - -### Q2 — Fast generation with camera control and multi-reference - -Q2 builds on Q1 with 3x faster inference, support for up to 7 reference images, enhanced subject consistency, and cinematic camera language controls including pan, zoom, and rotation. Ideal for iterative workflows and shot-level direction. Available as text-to-video, image-to-video, reference-to-video, and first-last-frame-to-video. - - - - Try the Q2 Text-to-Video workflow instantly on Comfy Cloud. - - - Download the workflow JSON. - - - - - - Try the Q2 Image-to-Video workflow instantly on Comfy Cloud. - - - Download the workflow JSON. - - - - - - Try the Q2 Reference-to-Video workflow instantly on Comfy Cloud. - - - Download the workflow JSON. - - - - - - Try the Q2 First-Last-Frame-to-Video workflow instantly on Comfy Cloud. - - - Download the workflow JSON. - - - -### Q3 — Latest generation for narrative and animation - -Vidu Q3 is the newest and most capable model generation. It delivers the highest visual quality with a particular strength in anime and 2D animation styles, making it the top choice for animation production. Q3 also introduces improved narrative understanding for more coherent storytelling. Generation is remarkably fast — a 4-second clip can be produced in as little as 10 seconds. Available as text-to-video and image-to-video. - - - - Try the Q3 Text-to-Video workflow instantly on Comfy Cloud. - - - Download the workflow JSON. - - - - - - Try the Q3 Image-to-Video workflow instantly on Comfy Cloud. - - - Download the workflow JSON. - - diff --git a/tutorials/partner-nodes/wavespeed/overview.mdx b/tutorials/partner-nodes/wavespeed/overview.mdx new file mode 100644 index 000000000..272e584ef --- /dev/null +++ b/tutorials/partner-nodes/wavespeed/overview.mdx @@ -0,0 +1,35 @@ +--- +title: "Image & Video Enhancement with WaveSpeed" +description: "Upscale video and images, and restore old photos using WaveSpeed AI models — FLSH-VSR video upscaling, SeedVR2 image restoration, and general image upscaling, all within ComfyUI." +sidebarTitle: "WaveSpeed" +--- + +import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; +import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; + +WaveSpeed is a high-performance AI media enhancement platform, now available as native workflow nodes in ComfyUI. It offers **image upscaling** and **SeedVR2 AI image fix** for photos, plus **FLSH-VSR video upscaling** — all powered by ComfyCloud with no local GPU needed. + +## Key capabilities + +- **FLSH-VSR video upscaling** — Temporal-coherent AI video super-resolution that maintains consistent detail across frames, ideal for archiving, AI video enhancement, and SD-to-HD conversion. +- **General image upscaling** — Multi-purpose image enlargement and quality improvement covering photographs, digital art, renders, and screenshots. +- **SeedVR2 AI image fix** — Diffusion-based restoration model for recovering damaged, degraded, or low-quality images, including old photo repair, noise and artifact removal, and lost detail reconstruction. +- **No local GPU required** — All processing runs on ComfyCloud's infrastructure, keeping your local hardware free for generation and composition. +- **Easy pipeline integration** — Simple drop-in nodes that work as a final enhancement step or intermediate preprocessing stage in any ComfyUI workflow. +- **Broad content compatibility** — Models handle diverse input types including photographs, AI art, video frames, and compressed web images. + + + + +## Models + +WaveSpeed offers two model groups in ComfyUI. The **Upscale** group covers general image upscaling and SeedVR2 restoration, while **FlashVSR** provides temporal-aware video super-resolution. + + + + General image upscaling and SeedVR2 AI image restoration. + + + Temporal-coherent AI video upscaling. + + diff --git a/tutorials/partner-nodes/wavespeed/wavespeed-flashvsr.mdx b/tutorials/partner-nodes/wavespeed/wavespeed-flashvsr.mdx new file mode 100644 index 000000000..e940431ef --- /dev/null +++ b/tutorials/partner-nodes/wavespeed/wavespeed-flashvsr.mdx @@ -0,0 +1,27 @@ +--- +title: "WaveSpeed FlashVSR" +description: "Upscale low-resolution video to higher quality using WaveSpeed's temporal-aware FLSH-VSR super-resolution model in ComfyUI." +sidebarTitle: "WaveSpeed FlashVSR" +--- + +import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; +import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; + +WaveSpeed FlashVSR (FLSH-VSR) is a temporal-coherent AI video super-resolution model that upscales low-resolution video while maintaining consistent detail across frames. Ideal for archiving, AI video enhancement, and SD-to-HD conversion. + +## Key capabilities + +- **Temporal-coherent upscaling** — Maintains frame-to-frame consistency for smooth, realistic output. +- **SD-to-HD conversion** — Upscales standard-definition video to high definition with realistic detail. +- **AI video enhancement** — Ideal for archiving vintage footage and improving compressed video quality. +- **No local GPU required** — All processing runs on ComfyCloud's infrastructure. + + + + +## Available workflows + + + Launch the FLSH-VSR video upscale workflow on ComfyCloud + Download the workflow JSON for local ComfyUI + diff --git a/tutorials/partner-nodes/wavespeed/wavespeed-upscale.mdx b/tutorials/partner-nodes/wavespeed/wavespeed-upscale.mdx new file mode 100644 index 000000000..62a7e26a9 --- /dev/null +++ b/tutorials/partner-nodes/wavespeed/wavespeed-upscale.mdx @@ -0,0 +1,40 @@ +--- +title: "WaveSpeed Upscale" +description: "Upscale images and restore old photos using WaveSpeed AI — general image upscaling and SeedVR2 AI image restoration in ComfyUI." +sidebarTitle: "WaveSpeed Upscale" +--- + +import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; +import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; + +WaveSpeed Upscale covers two image enhancement workflows: **general image upscaling** for multi-purpose enlargement and quality improvement, and **SeedVR2 AI image fix** for diffusion-based restoration of damaged, degraded, or low-quality photos. + +## Key capabilities + +- **General image upscaling** — Multi-purpose image enlargement and quality improvement for photographs, digital art, renders, and screenshots. +- **SeedVR2 AI image fix** — Diffusion-based restoration for recovering damaged, degraded, or low-quality images, including old photo repair and artifact removal. +- **No local GPU required** — All processing runs on ComfyCloud's infrastructure. +- **Broad content compatibility** — Handles photographs, AI art, video frames, and compressed web images. + + + + +## Available workflows + +### Image Upscaling + +General-purpose AI image upscaling for photographs, renders, and digital artwork. Enlarges and enhances while preserving natural detail. + + + Launch the image upscale workflow on ComfyCloud + Download the workflow JSON for local ComfyUI + + +### SeedVR2 AI Image Fix + +Restore damaged, degraded, or low-quality images using WaveSpeed's diffusion-based restoration model. Repairs noise, artifacts, color degradation, and missing facial details. + + + Launch the SeedVR2 image fix workflow on ComfyCloud + Download the workflow JSON for local ComfyUI + diff --git a/tutorials/partner-nodes/wavespeed/wavespeed.mdx b/tutorials/partner-nodes/wavespeed/wavespeed.mdx deleted file mode 100644 index 322bba19e..000000000 --- a/tutorials/partner-nodes/wavespeed/wavespeed.mdx +++ /dev/null @@ -1,51 +0,0 @@ ---- -title: "Image & Video Enhancement with WaveSpeed" -description: "Upscale video and images, and restore old photos using WaveSpeed AI models — FLSH-VSR video upscaling, SeedVR2 image restoration, and general image upscaling, all within ComfyUI." -sidebarTitle: "WaveSpeed" ---- - -import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; -import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; - -WaveSpeed is a high-performance AI media enhancement platform, now available as native workflow nodes in ComfyUI. It offers **FLSH-VSR video upscaling**, **general image upscaling**, and the **SeedVR2 AI image fix** for restoring damaged photos — all powered by ComfyCloud with no local GPU needed. - -## Key capabilities - -- **FLSH-VSR video upscaling** — Temporal-coherent AI video super-resolution that maintains consistent detail across frames, ideal for archiving, AI video enhancement, and SD-to-HD conversion. -- **General image upscaling** — Multi-purpose image enlargement and quality improvement covering photographs, digital art, renders, and screenshots. -- **SeedVR2 AI image fix** — Diffusion-based restoration model for recovering damaged, degraded, or low-quality images, including old photo repair, noise and artifact removal, and lost detail reconstruction. -- **No local GPU required** — All processing runs on ComfyCloud's infrastructure, keeping your local hardware free for generation and composition. -- **Easy pipeline integration** — Simple drop-in nodes that work as a final enhancement step or intermediate preprocessing stage in any ComfyUI workflow. -- **Broad content compatibility** — Models handle diverse input types including photographs, AI art, video frames, and compressed web images. - - - - -## Available workflows - -### FLSH-VSR Video Upscaling - -Upscale low-resolution video to higher quality using WaveSpeed's temporal-aware super-resolution model. Maintains frame-to-frame consistency for smooth, realistic output. - - - Launch the FLSH-VSR video upscale workflow on ComfyCloud - Download the workflow JSON for local ComfyUI - - -### Image Upscaling - -General-purpose AI image upscaling for photographs, renders, and digital artwork. Enlarges and enhances while preserving natural detail. - - - Launch the image upscale workflow on ComfyCloud - Download the workflow JSON for local ComfyUI - - -### SeedVR2 AI Image Fix - -Restore damaged, degraded, or low-quality images using WaveSpeed's diffusion-based restoration model. Repairs noise, artifacts, color degradation, and missing facial details. - - - Launch the SeedVR2 image fix workflow on ComfyCloud - Download the workflow JSON for local ComfyUI - From d5c294519ad508546f5f68fecea3c28ef1f2bc4e Mon Sep 17 00:00:00 2001 From: linmoumou Date: Sat, 2 May 2026 15:10:41 +0800 Subject: [PATCH 6/9] Update Grok docs: add tier info (Pro/Standard/Beta), generation speed, resolution support, style specialties - Remove 'beta release' phrasing, reflect Pro/Standard/Beta coexistence - Add ~4s/image speed benchmark, 1K/2K resolution info, image edit input limits - Add video generation time estimates (~60s extend, ~100s 720p reference) - Add style specialties: anime, cyberpunk, motion graphics, world rebuilding --- tutorials/partner-nodes/grok/grok-imagine.mdx | 12 ++++++++---- tutorials/partner-nodes/grok/grok-video.mdx | 12 ++++++++---- tutorials/partner-nodes/grok/overview.mdx | 12 +++++++----- 3 files changed, 23 insertions(+), 13 deletions(-) diff --git a/tutorials/partner-nodes/grok/grok-imagine.mdx b/tutorials/partner-nodes/grok/grok-imagine.mdx index 815f9cfe2..c390a7319 100644 --- a/tutorials/partner-nodes/grok/grok-imagine.mdx +++ b/tutorials/partner-nodes/grok/grok-imagine.mdx @@ -1,19 +1,23 @@ --- title: "Grok Imagine — Text to Image & Image Editing" -description: "Generate and edit images with xAI's Grok Imagine models — cinematic quality with dramatic lighting and rich color saturation." +description: "Generate and edit images with xAI's Grok Imagine models — Pro, Standard, and Beta tiers, fast ~4s generation, cinematic quality with dramatic lighting and rich color saturation." sidebarTitle: "Grok Imagine" --- import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; -Generate images from text prompts and edit existing images using Grok Imagine by xAI. With pro/standard/beta quality tiers, Grok delivers a distinct moody, cinematic aesthetic — dramatic lighting, rich color saturation, and strong character rendering with particular strengths in anime and portraiture. +Generate images from text prompts and edit existing images using Grok Imagine by xAI, available in **Pro**, **Standard**, and **Beta** tiers. Grok delivers a distinct moody, cinematic aesthetic — dramatic lighting, rich color saturation, and strong character rendering with particular strengths in anime, portraiture, and cyberpunk styles. Image generation is notably fast at approximately **4 seconds per image**, making it well-suited for rapid iteration and exploration. + +Available output resolutions: **1K** and **2K**. Image editing supports 1 input image for Pro, or up to 3 input images for Standard/Beta. ## Key capabilities -- **Text to Image** — Generate images from text prompts with three quality tiers -- **Image Editing** — Modify existing images using natural language instructions +- **Fast text-to-image generation** — ~4 seconds per image, ideal for rapid iteration +- **Image Editing** — Modify existing images with flexible aspect ratio options - **Cinematic aesthetic** — Dramatic lighting, rich color saturation, moody atmosphere +- **Strong character rendering** — Excellent facial consistency, expressive lighting and framing +- **Anime & cyberpunk** — Stylized illustrations and neon compositions translate especially well diff --git a/tutorials/partner-nodes/grok/grok-video.mdx b/tutorials/partner-nodes/grok/grok-video.mdx index ae36292a6..d03e355d0 100644 --- a/tutorials/partner-nodes/grok/grok-video.mdx +++ b/tutorials/partner-nodes/grok/grok-video.mdx @@ -7,14 +7,18 @@ sidebarTitle: "Grok Video" import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; -Generate video from text descriptions, edit existing video footage, extend clips naturally, and create consistent video using reference images — all with Grok's signature moody visual style. Grok Video brings xAI's cinematic aesthetic to motion content. +Generate video from text descriptions, edit existing video footage, extend clips naturally, and create consistent video using reference images — all with Grok's signature moody visual style. Grok Video brings xAI's cinematic aesthetic to motion content. The model is particularly effective at motion graphics, 2D illustrations, and image-to-video world rebuilding from new angles. + +Available in Standard and Beta tiers. ## Key capabilities - **Text to Video** — Generate video from text descriptions with cinematic quality -- **Video Editing** — Edit existing video footage with text instructions -- **Video Extending** — Extend an input video by up to 10 seconds with natural continuations -- **Reference to Video** — Generate video guided by up to 7 reference images for identity consistency +- **Video Editing** — Edit existing video footage with text instructions (input: 1-8.7s, max 50MB) +- **Video Extending** — Extend an input video of up to 15 seconds by up to 10 seconds with natural continuations (~60s generation time for 10s extension) +- **Reference to Video** — Generate video guided by up to 7 reference images for identity consistency (~100s for 10s video at 720p) +- **Motion graphics & illustrations** — Flat and graphic inputs animate cleanly into bold video results +- **Image-to-video world rebuilding** — Reconstructs scenes from new angles or perspectives diff --git a/tutorials/partner-nodes/grok/overview.mdx b/tutorials/partner-nodes/grok/overview.mdx index b18738e3e..d6237f2e6 100644 --- a/tutorials/partner-nodes/grok/overview.mdx +++ b/tutorials/partner-nodes/grok/overview.mdx @@ -1,22 +1,24 @@ --- title: "Grok Image & Video Generation" -description: "Generate images and videos with xAI's Grok models — fast, moody aesthetics with strong anime and cinematic character rendering." +description: "Generate images and videos with xAI's Grok models — fast ~4s image generation, moody cinematic aesthetics, with strong anime and character rendering. Available in Pro, Standard, and Beta tiers." sidebarTitle: "Grok" --- import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; -Grok Imagine by xAI brings a distinct, moody visual style to AI image and video generation. It produces cinematic-quality images with dramatic lighting and rich color saturation, and supports text-to-video, video editing, and reference-to-video generation with consistent character identity. +Grok Imagine by xAI brings a distinct, moody visual style to AI image and video generation. It produces cinematic-quality images with dramatic lighting and rich color saturation, and supports text-to-video, video editing, and reference-to-video generation with consistent character identity. Image generation is notably fast at approximately 4 seconds per image, ideal for rapid iteration. + +Grok performs particularly well with anime, cyberpunk, and cinematic portrait styles, and excels at reconstructing scenes from new angles in image-to-video workflows. ## Key capabilities -- **Text to Image** — Generate images from text prompts with pro/standard/beta quality tiers +- **Text to Image** — Generate images from text prompts (~4 seconds per image) - **Image Editing** — Edit existing images with text instructions - **Text to Video** — Generate video from text descriptions - **Video Editing** — Edit existing video footage -- **Video Extending** — Extend an input video (15s in, extend up to 10s) -- **Reference to Video** — Generate consistent video using up to 7 reference images +- **Video Extending** — Extend an input video (15s in, extend up to 10s, ~60s generation time) +- **Reference to Video** — Generate consistent video using up to 7 reference images (~100s for 10s at 720p) From 5a50179dacca7c096e0dc93332ceb341497cdf5c Mon Sep 17 00:00:00 2001 From: linmoumou Date: Sat, 2 May 2026 22:44:34 +0800 Subject: [PATCH 7/9] Update Vidu Q2 docs: add camera language details, expressions, multi-character coordination, dynamic rendering improvements MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - Expand camera control details (push/pull/orbit/follow/close-up) - Add expression & micro-movement, dynamic rendering capabilities - Add multi-character coordination and scene composition - Note Turbo/Pro/Standard model variants - Remove 'Four generation modes' → 'Six generation modes' --- tutorials/partner-nodes/vidu/vidu-q2.mdx | 26 +++++++++++++++++++----- 1 file changed, 21 insertions(+), 5 deletions(-) diff --git a/tutorials/partner-nodes/vidu/vidu-q2.mdx b/tutorials/partner-nodes/vidu/vidu-q2.mdx index 6e5fbf46b..d364c775c 100644 --- a/tutorials/partner-nodes/vidu/vidu-q2.mdx +++ b/tutorials/partner-nodes/vidu/vidu-q2.mdx @@ -1,20 +1,25 @@ --- title: "Vidu Q2 - Fast generation with camera control" -description: "Generate video 3x faster with multi-reference support and cinematic camera controls using Vidu Q2 in ComfyUI" +description: "Generate video 3x faster with multi-reference support, cinematic camera controls, and enhanced dynamic rendering using Vidu Q2 in ComfyUI" sidebarTitle: "Vidu Q2" --- import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; -Vidu Q2 builds on the Q1 foundation with significantly faster inference (3x speedup), support for up to 7 reference images, enhanced subject consistency, and cinematic camera language controls including pan, zoom, and rotation. It is available as text-to-video, image-to-video, reference-to-video, and first-last-frame-to-video, making it ideal for iterative workflows and shot-level direction. +Vidu Q2 builds on the Q1 foundation with significantly faster inference (3x speedup), support for up to 7 reference images, enhanced subject consistency, and cinematic camera language controls including push, pull, orbit, follow, and close-up strategies. It also delivers higher-fidelity dynamic rendering with smoother large motions and more believable physical feedback, along with finer facial expressions, eye movement, and subtle gestures for more expressive characters. + +Available in Turbo, Pro, and Standard model variants. ## Key capabilities - **3x faster inference** — Significantly reduced generation time compared to Q1 -- **Multi-reference support** — Up to 7 reference images for enhanced subject consistency -- **Cinematic camera control** — Pan, zoom, rotation, and other camera language for precise shot direction -- **Four generation modes** — Text-to-video, image-to-video, reference-to-video, and first-last-frame-to-video +- **Multi-image reference** — Up to 7 reference images ensuring identity, scene, and narrative consistency across subjects +- **Cinematic camera control** — Push, pull, orbit, follow, close-up and other camera language for precise shot direction +- **High-fidelity dynamic rendering** — Smoother large motions with more believable physical feedback +- **Expressions & micro-movements** — Finer facial expressions, eye movement, and subtle gestures for expressive characters +- **Multi-character coordination** — Stable character interaction and complex scene composition across shots +- **Six generation modes** — Text-to-video, image-to-video, reference-to-video, first-last-frame-to-video, video extension, and multi-frame generation @@ -64,3 +69,14 @@ Vidu Q2 builds on the Q1 foundation with significantly faster inference (3x spee Download the workflow JSON. + +### First last frame to video + + + + Try the Q2 First-Last-Frame-to-Video workflow instantly on Comfy Cloud. + + + Download the workflow JSON. + + From f3555b50fb4956dcc105c7a89c7ea6fe75a042d8 Mon Sep 17 00:00:00 2001 From: linmoumou Date: Sat, 2 May 2026 22:48:01 +0800 Subject: [PATCH 8/9] Update Vidu Q3 docs: add Pro/Turbo tiers, audio support, 2K resolution, duration/amplitude details Borrowed context from wavespeed.ai model page for parameter details. --- tutorials/partner-nodes/vidu/vidu-q3.mdx | 15 +++++++++++---- 1 file changed, 11 insertions(+), 4 deletions(-) diff --git a/tutorials/partner-nodes/vidu/vidu-q3.mdx b/tutorials/partner-nodes/vidu/vidu-q3.mdx index 54ff4ede9..2ae672b0b 100644 --- a/tutorials/partner-nodes/vidu/vidu-q3.mdx +++ b/tutorials/partner-nodes/vidu/vidu-q3.mdx @@ -1,20 +1,25 @@ --- title: "Vidu Q3 - Latest generation for narrative and animation" -description: "Generate high-quality video with improved narrative understanding and outstanding anime output using Vidu Q3 in ComfyUI" +description: "Generate high-quality video with improved narrative understanding, outstanding anime output, and audio support using Vidu Q3 in ComfyUI" sidebarTitle: "Vidu Q3" --- import ReqHint from "/snippets/tutorials/partner-nodes/req-hint.mdx"; import UpdateReminder from "/snippets/tutorials/update-reminder.mdx"; -Vidu Q3 is the newest and most capable Vidu model generation, powered by the Vidu 3.5 engine. It delivers the highest visual quality with a particular strength in anime and 2D animation styles, making it the top choice for animation production pipelines. Q3 also introduces improved narrative understanding for more coherent storytelling. Generation is remarkably fast — a 4-second clip can be produced in as little as 10 seconds. +Vidu Q3 is the newest and most capable Vidu model generation, powered by the Vidu 3.5 engine. It delivers the highest visual quality with a particular strength in anime and 2D animation styles, making it the top choice for animation production pipelines. Q3 also supports synchronized audio generation including dialogue, sound effects, and background music. + +Available in **Pro** (up to 2K resolution) and **Turbo** (optimized for speed) tiers. ## Key capabilities - **Vidu 3.5 engine** — The latest engine delivering the highest visual quality across the Vidu family - **Exceptional anime quality** — Industry-leading output for 2D animation production -- **Improved narrative understanding** — More coherent storytelling across generated clips -- **Fast generation speed** — 4-second clips in as fast as 10 seconds +- **Synchronized audio** — Generate dialogue, sound effects, and background music matching the scene +- **Multi-resolution support** — 720p, 1080p on all models, plus 2K on Pro for image-to-video +- **Flexible duration** — Create videos from 1 to 16 seconds +- **Motion amplitude control** — Adjust movement intensity from subtle to dynamic +- **Three generation modes** — Text-to-video, image-to-video, and start/end-frame-to-video @@ -42,3 +47,5 @@ Vidu Q3 is the newest and most capable Vidu model generation, powered by the Vid Download the workflow JSON. + + From c31b58f07825ade79049468acb5da08607617cc3 Mon Sep 17 00:00:00 2001 From: linmoumou Date: Sun, 3 May 2026 07:10:26 +0800 Subject: [PATCH 9/9] =?UTF-8?q?Fix=20ERNIE-Image-Turbo=20diffusion=20model?= =?UTF-8?q?=20link=20(ernie-image=20=E2=86=92=20ernie-image-turbo)?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The ERNIE-Image-Turbo section's first model download Card pointed to ernie-image.safetensors (the base model) instead of ernie-image-turbo.safetensors. Updated the Card href, title, and the corresponding path in the Turbo section's model storage code block. Fixes #961 --- tutorials/image/ernie-image/ernie-image.mdx | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/tutorials/image/ernie-image/ernie-image.mdx b/tutorials/image/ernie-image/ernie-image.mdx index 1192bd8a6..45f90724f 100644 --- a/tutorials/image/ernie-image/ernie-image.mdx +++ b/tutorials/image/ernie-image/ernie-image.mdx @@ -92,7 +92,7 @@ You can find all repackaged model files at [Comfy-Org/ERNIE-Image](https://huggi ### ERNIE-Image-Turbo model downloads - + Diffusion model for ERNIE-Image-Turbo. @@ -112,7 +112,7 @@ You can find all repackaged model files at [Comfy-Org/ERNIE-Image](https://huggi 📂 ComfyUI/ ├── 📂 models/ │ ├── 📂 diffusion_models/ -│ │ └── ernie-image.safetensors +│ │ └── ernie-image-turbo.safetensors │ ├── 📂 text_encoders/ │ │ ├── ministral-3-3b.safetensors │ │ └── ernie-image-prompt-enhancer.safetensors