Provides AI vision analysis using Google Gemini models, enabling image understanding, OCR, error diagnosis, chart analysis, and UI design evaluation.
Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@mcp-vision-analyze What's shown in this screenshot?"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
A lightweight Model Context Protocol (MCP) server that provides AI vision analysis capabilities. Analyze images using Google's Gemini models via OpenRouter.
Works with: Claude Desktop, Claude Code CLI, Antigravity, Cursor, Pi Agent, Windsurf, Cline, VS Code, and any MCP-compatible client.
🖼️ Analyze images from local file paths or HTTP(S) URLs
🧠 Powered by Google Gemini models (2.5 Flash Lite, 3 Flash Preview, etc.)
💰 Ultra cheap — ~$0.001 per image with Gemini 2.5 Flash Lite
📝 Extract text from screenshots (OCR)
🐛 Diagnose error messages in screenshots
📊 Analyze charts, diagrams, and UI designs
🔒 Secure — API key stays local, never sent anywhere except OpenRouter
Related MCP server: OpenRouter Image MCP Server
Model | Input | Output | Best For |
| $0.10/M tokens | $0.40/M tokens | Default — cheapest |
| $0.25/M tokens | $1.50/M tokens | Better quality |
| $0.50/M tokens | $3.00/M tokens | Best reasoning |
| $0.38/M tokens | $1.88/M tokens | Latest model |
💡 1 image analysis ≈ 1,300 input tokens + 150 output tokens ≈ $0.001
Sign up at OpenRouter and get your API key from openrouter.ai/keys.
# Option A: Use directly with npx (recommended — no install needed)
npx mcp-vision-analyze
# Option B: Clone and install manually
git clone https://github.com/rezkycodes/mcp-vision-analyze.git
cd mcp-vision-analyze
npm installclaude_desktop_config.json){
"mcpServers": {
"vision-analyze": {
"command": "npx",
"args": ["-y", "mcp-vision-analyze"],
"env": {
"OPENROUTER_API_KEY": "sk-or-v1-your-key-here"
}
}
}
}~/.pi/agent/mcp.json){
"vision-analyze": {
"transport": "stdio",
"command": "node",
"args": ["/absolute/path/to/mcp-vision-analyze/index.js"],
"env": {
"OPENROUTER_API_KEY": "sk-or-v1-your-key-here"
},
"directTools": true
}
}# Option 1: Using the CLI command
claude mcp add vision-analyze \
-e OPENROUTER_API_KEY=sk-or-v1-your-key-here \
-- npx -y mcp-vision-analyze
# Option 2: Using JSON config
claude mcp add-json vision-analyze '{
"command": "npx",
"args": ["-y", "mcp-vision-analyze"],
"env": {
"OPENROUTER_API_KEY": "sk-or-v1-your-key-here"
}
}'Or add to .mcp.json in your project root:
{
"mcpServers": {
"vision-analyze": {
"command": "npx",
"args": ["-y", "mcp-vision-analyze"],
"env": {
"OPENROUTER_API_KEY": "sk-or-v1-your-key-here"
}
}
}
}Add to Antigravity Settings → MCP, or edit ~/.config/Antigravity/User/globalStorage/*/mcp.json:
{
"mcpServers": {
"vision-analyze": {
"command": "npx",
"args": ["-y", "mcp-vision-analyze"],
"env": {
"OPENROUTER_API_KEY": "sk-or-v1-your-key-here"
}
}
}
}.cursor/mcp.json){
"mcpServers": {
"vision-analyze": {
"command": "npx",
"args": ["-y", "mcp-vision-analyze"],
"env": {
"OPENROUTER_API_KEY": "sk-or-v1-your-key-here"
}
}
}
}~/.config/Code/User/mcp.json){
"servers": {
"vision-analyze": {
"type": "stdio",
"command": "npx",
"args": ["-y", "mcp-vision-analyze"],
"env": {
"OPENROUTER_API_KEY": "sk-or-v1-your-key-here"
}
}
}
}Add the server config with:
command: npx
args: ["-y", "mcp-vision-analyze"]
env: { "OPENROUTER_API_KEY": "your-key" }
Once configured, the vision_analyze tool becomes available in your MCP client:
{
"image_url": "/path/to/screenshot.png",
"prompt": "What is shown in this screenshot?"
}{
"image_url": "https://example.com/image.jpg",
"prompt": "Describe the architectural style of this building"
}{
"image_url": "/path/to/photo.png",
"prompt": "Extract all text from this image"
}{
"image_url": "/path/to/error-screenshot.png",
"prompt": "What error is shown and how to fix it?"
}{
"image_url": "/path/to/image.png",
"prompt": "Analyze this chart",
"model": "google/gemini-3-flash-preview"
}mcp-vision-analyze/
├── index.js # MCP server main file
├── package.json # Node.js dependencies
├── .env.example # Environment variables template
├── .env # Your API key (git-ignored)
├── .gitignore # Git ignore rules
└── README.md # This fileVariable | Required | Default | Description |
| ✅ Yes | — | Your OpenRouter API key |
| No |
| Gemini model to use |
JPEG / JPG
PNG
GIF
WebP
BMP (auto-converted to PNG)
SVG (auto-rasterized to PNG)
The MCP server itself is free and open-source. The Gemini models on OpenRouter are paid but very cheap (~$0.001 per image).
Currently supports OpenRouter models. You can change VISION_MODEL to any model available on OpenRouter that supports vision (GPT-4o, Claude, etc.).
No. It requires an internet connection to call the OpenRouter API.
No. Images are processed in-memory and sent directly to OpenRouter's API. Nothing is stored on disk (except temporary files that are auto-cleaned).
MIT
Inspired by the vision_analyze tool from Hermes Agent.
Unclaimed servers have limited discoverability.
If you are the server author, to access and configure the admin panel.
LLM chat, text summarization and AI image generation
Analyze images from multiple angles to extract detailed insights or quick summaries. Describe visu…
OCR, transcription, file extraction, and image generation for AI agents via MCP.
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/rezkycodes/mcp-vision-analyze'
If you have feedback or need assistance with the MCP directory API, please join our Discord server