🤖 OllamaMulti-RAG: Neura-Nix

An AI-powered multimodal assistant that seamlessly understands Text, Images, PDFs, and Audio — built to boost productivity, empower businesses, and transform everyday workflows.

🌟 Introduction

Neura-Nix is a next-gen multimodal assistant that combines the power of Ollama, OpenAI, Whisper, and Redis into a single streamlined platform.
It enables natural and intelligent interaction across different mediums — chat with documents, analyze images, transcribe audio, and converse in real-time.

Built with Streamlit, Docker, and Redis caching, Neura-Nix is designed for speed, extensibility, and scalability.
Whether you’re an enterprise optimizing workflows or an individual boosting productivity, Neura-Nix adapts seamlessly.

🚀 Key Features

• 📖 Document Chat (PDF RAG) → Upload PDFs and extract context-aware insights.
• 🖼️ Image Analysis → Unlock hidden stories from visual data.
• 🎤 Audio to Text → Record or upload audio and let Whisper transcribe in seconds.
• 💬 Persistent Chat Sessions → Store, manage, and reload previous conversations.
• ⚡ Redis Caching → Lightning-fast performance with session-level caching.
• 🎯 Model Flexibility → Switch easily between Ollama and OpenAI models.
• 🔐 Secure by Design → Environment variables, .env.sample, and Dockerized deployment.
• 📊 Optimized Vector Storage → ChromaDB-backed semantic search for documents.
• 🎨 Modern UI → Built with Streamlit, responsive and minimal.

📽️ Demo

Coming soon... 🎬

Important

Use the Docker image or run the project locally via localhost to get started. Before proceeding, please contact me at Mail so I can share the credentials instead Docker extract your whole memory.

🛠️ Technology Stack

• Frontend/UI → Streamlit
• LLMs & Embeddings → Ollama + OpenAI
• Audio Transcription → Whisper
• Vector Database → ChromaDB
• Caching & Session Store → Redis Cloud
• Containerization → Docker + Docker Compose
• Web Server → NGINX (reverse proxy & static serving)
• Database → SQLite (lightweight local DB for session caching)
• Orchestration → GitHub Actions + Dependabot for CI/CD

Getting Started ⚙️

Prerequisites

Before setting up Neura-Nix, make sure you have the following installed:

• Python 3.10+ → Required for running the backend (tested on 3.10.12).

• Docker & Docker Compose → Preferred method for containerized deployment. Install Docker

• Git → For cloning and managing the repository. Install Git.

• Redis (Cloud or Local) → Used for caching and optimizing performance.
  Sign up for Redis Cloud or run a local instance.

• Ollama → Required for running local multimodal models.

  • Download Ollama Desktop (Windows/macOS)
  • Or install manually on Linux

• (Optional) GPU Support → If available, install NVIDIA drivers + CUDA toolkit for accelerated model performance.

⚡ Tip: Ensure environment variables (like REDIS_HOST, REDIS_PORT, OPENAI_API_KEY) are properly configured in your .env file before running the project.

Installation 🛠️

You can follow the official setup guide for Linux, Windows, and Docker below to run Neura-Nix {OllamaMulti-RAG} locally.

• First Read this License & their terms then proceed.
• Star ⭐ the Repository
• Fork the repository (Optional)
• Project Setup:

1. Clone the repository:

    git clone https://github.com/UjjwalSaini07/OllamaMulti-RAG.git

2. Navigate to the project main directory:

    cd NexGen-Quillix

Important

All these cd directory paths are relative to the root directory of the cloned project.

After Cloning the repository and choose your preferred installation method.

🔹 Method 1: Docker Compose {Not Fastest Step}

1. Modify Docker Compose →

   • Remove docker-compose.yml.
   • Rename docker-compose_with_ollama.yml → docker-compose.yml.

2. Set model save path → Update line 25 in the docker-compose.yml file.

3. Run Neura-Nix

     docker compose up

   ⚡ If you don’t have a GPU → remove the deploy section from the compose file.

4. Optional Configurations

   • Edit config.yaml to match your needs.
   • Add custom icons → replace user_image.png and/or bot_image.png inside the chat_icons folder.

5. Access the app → Open http://0.0.0.0:8501 in your browser.

6. Pull Models → Visit Ollama Library and pull models:

     /pull MODEL_NAME

   ✅ You need:

   • An embedding model → e.g., nomic-embed-text for PDFs.
   • An image-capable model → e.g., llava for image analysis.

🔹 Method 2: Windows (Best Performance)

⚠️ Using Ollama inside Docker on Windows can be slow → prefer local installation.

1. Install Ollama Desktop.

2. Update config → In config.yaml, use line 4 (Windows) for base_url, remove line 3 {Default is Correct}.

3. Start Neura-Nix Up the Docker Container:

     docker compose up

4. Access → http://0.0.0.0:8501.

5. Pull models as prescribed in Method 1.

🔹 Method 3: Manual Install

1. Install Ollama.
2. Create Python venv (tested with Python 3.10.12).
3. Install dependencies:

     pip install --upgrade pip
     pip install -r requirements.txt
     pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cpu

4. Run setup:

     python3 database_operations.py   # initialize SQLite DB
     streamlit run app.py

5. Pull models (embedding + multimodal as above).
6. Optional → update config.yaml and add custom chat icons.

📊 Business Optimization

Neura-Nix is not just a research tool — it’s built to optimize workflows and unlock ROI:

• 🏢 Enterprise Teams → Automate document review, compliance checks, and data-heavy workflows.
• 📈 Startups → Accelerate content creation, customer support, and product research.
• 👨‍💻 Freelancers & Creators → Boost productivity with multimodal AI (chat, docs, media).
• 🔒 Privacy by Design → Keep your sensitive data secure with local-first deployment.

📚 Documentation

To help you navigate and extend Neura-Nix, we’ve structured the documentation into multiple layers:

• Python Documentation
• Streamlit Documentation
• Ollama Documentation
• OpenAI API Documentation
• Whisper Models (OpenAI)
• Redis Documentation
• ChromaDB Documentation
• Docker Documentation
• Nginx Documentation
• GitHub Actions Documentation

Author ✍️

• @Ujjwal Saini

Contact 📞

Feel free to reach out if you have any questions or suggestions!

• Raise an issue for the same Issue
• Github: @Ujjwal Saini
• Mail: Mail ID

License 📄

License Credential Check.
You can use this project the way you want. Feel free to credit me if you want to!

Feedback and Contributions 💌

Feedback and contributions are always welcome! Feel free to open an Issue.