From bfec852c9272a85c020c0f677b986537278451fb Mon Sep 17 00:00:00 2001
From: NickRivers1983 <nick@example.com>
Date: Thu, 14 May 2026 23:47:19 -0500
Subject: [PATCH] docs: add opencode-vision MCP server example

---
 packages/web/src/content/docs/mcp-servers.mdx | 46 +++++++++++++++++++
 1 file changed, 46 insertions(+)

diff --git a/packages/web/src/content/docs/mcp-servers.mdx b/packages/web/src/content/docs/mcp-servers.mdx
index 1b3006b1cbf2..34d875d7b935 100644
--- a/packages/web/src/content/docs/mcp-servers.mdx
+++ b/packages/web/src/content/docs/mcp-servers.mdx
@@ -509,3 +509,49 @@ Alternatively, you can add something like this to your [AGENTS.md](/docs/rules/)
 ```md title="AGENTS.md"
 If you are unsure how to do something, use `gh_grep` to search code examples from GitHub.
 ```
+
+---
+
+### opencode-vision
+
+Add vision capabilities to **any** text-only model. When your model can't process images natively (e.g., big-pickle, DeepSeek, MiMo), this MCP server handles image analysis via **Google Gemini Vision API** (FREE tier) and **local tesseract OCR**, returning text descriptions that any model can understand.
+
+**Requirements:**
+- Python >= 3.10
+- Google Gemini API key (get one free at [aistudio.google.com](https://aistudio.google.com/))
+
+```json title="opencode.json"
+{
+  "$schema": "https://opencode.ai/config.json",
+  "mcp": {
+    "vision": {
+      "type": "local",
+      "command": ["python3", "-m", "opencode_vision.server"],
+      "enabled": true,
+      "timeout": 30000
+    }
+  }
+}
+```
+
+**Setup:**
+
+1. Install the server:
+   ```bash
+   pip install opencode-vision
+   ```
+2. Set your Gemini API key in `~/.config/opencode/.env`:
+   ```bash
+   echo 'GOOGLE_API_KEY=your_key_here' >> ~/.config/opencode/.env
+   ```
+3. Restart OpenCode.
+
+**Available tools:**
+
+| Tool | Description |
+|------|-------------|
+| `vision_describe(path, prompt?)` | Describe an image's composition, colors, text, and context |
+| `vision_ocr(path)` | Extract text from images via tesseract + Gemini fallback |
+| `vision_analyze(path)` | Full analysis: metadata + description + OCR |
+
+Learn more at [github.com/NickRivers1983/opencode-vision](https://github.com/NickRivers1983/opencode-vision) or `pip install opencode-vision`.