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., "@Vision MCP Server extract text from this image: receipt.jpg"
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.
Give your AI agent eyes. An MCP server providing multimodal vision capabilities — image analysis, OCR, image comparison, and video analysis — powered by any OpenAI-compatible vision model.
让你的 AI 代理拥有视觉能力。 通过任何 OpenAI 兼容的视觉模型，提供图像分析、OCR 文字识别、图像对比和视频分析。
Features · Quick Start · Tools · Models · 中文说明
Tool | Description |
🔍 | Analyze images with natural language prompts |
📝 | Extract text from images (plain text / Markdown / JSON) |
🔀 | Compare 2–4 images side by side |
🎬 | Analyze video content (requires video-capable model) |
Plus:
🌐 OpenAI-compatible — Works with any vision model via standard API
📁 Local files & URLs — Auto-converts local files to base64
⚙️ Configurable — Environment variables, config files, or both
Related MCP server: Vision MCP for Reasonix
git clone https://github.com/Loveacup/vision-mcp-server.git
cd vision-mcp-server
npm install && npm run buildCreate a .env file in the project root:
VISION_BASE_URL=http://your-server:port/v1/chat/completions
VISION_MODEL=Qwen3-VL-32B
VISION_API_KEY=your-api-key # optional for local models{
"baseUrl": "http://your-server:port/v1/chat/completions",
"model": "Qwen3-VL-32B",
"apiKey": "your-api-key",
"maxTokens": 4096,
"temperature": 0.7
}npm startThe server communicates over stdio, designed to be launched by an MCP client such as Claude Code.
Add to your ~/.mcp.json:
{
"mcpServers": {
"vision": {
"command": "node",
"args": ["/path/to/vision-mcp-server/dist/index.js"],
"env": {
"VISION_BASE_URL": "http://your-server:port/v1/chat/completions",
"VISION_MODEL": "Qwen3-VL-32B",
"VISION_API_KEY": "your-api-key"
}
}
}
}Replace
/path/to/vision-mcp-serverwith the actual install path.
Configuration priority: environment variables > config file > defaults
Variable | Config Key | Default | Description |
|
| (required) | OpenAI-compatible chat completions endpoint |
|
|
| Model name |
|
| (empty) | API key (optional for local models) |
|
|
| Max response tokens |
|
|
| Sampling temperature |
analyze_imageAnalyze an image with a vision language model.
Parameter | Type | Required | Default | Description |
| string | ✅ | — | Local file path or URL |
| string |
| Analysis prompt | |
|
|
| Detail level |
ocr_imageExtract text from an image using OCR.
Parameter | Type | Required | Default | Description |
| string | ✅ | — | Local file path or URL |
| string |
| Language hint, e.g. | |
|
|
| Output format |
compare_imagesCompare 2–4 images and describe differences/similarities.
Parameter | Type | Required | Default | Description |
| string[] | ✅ | — | 2–4 image sources |
| string |
| Comparison prompt |
analyze_videoAnalyze video content. Requires a model with video support (e.g., Qwen3-VL).
Parameter | Type | Required | Default | Description |
| string | ✅ | — | Local file path or URL |
| string |
| Analysis prompt |
Model | Provider | Image | Video | Notes |
Qwen3-VL | Self-hosted / API | ✅ | ✅ | Recommended. Full multimodal support |
GPT-4o | OpenAI | ✅ | ❌ | Strong image analysis |
LLaVA | Self-hosted | ✅ | ❌ | Open-source alternative |
InternVL | Self-hosted | ✅ | ⚠️ | Strong multilingual OCR |
Any model served via vLLM, Ollama, LMDeploy, or other OpenAI-compatible servers should work.
Supported formats: JPEG, PNG, GIF, WebP, BMP, SVG | MP4, AVI, MOV, MKV, WebM
vision-mcp-server/
├── src/
│ ├── index.ts # MCP server entry point
│ ├── config.ts # Configuration loader
│ ├── types.ts # TypeScript type definitions
│ ├── tools/
│ │ ├── analyze-image.ts
│ │ ├── ocr-image.ts
│ │ ├── compare-images.ts
│ │ └── analyze-video.ts
│ └── utils/
│ ├── api-client.ts # OpenAI-compatible API client
│ └── file-handler.ts # Local file → base64
├── package.json
├── tsconfig.json
├── .env.example
└── LICENSEanalyze_image — 使用视觉语言模型分析图像，支持自然语言提问
ocr_image — OCR 文字识别，支持纯文本、Markdown、JSON 输出
compare_images — 对比 2–4 张图像，识别差异和相似之处
analyze_video — 分析视频内容（需要 Qwen3-VL 等支持视频的模型）
git clone https://github.com/Loveacup/vision-mcp-server.git
cd vision-mcp-server
npm install && npm run build配置 .env：
VISION_BASE_URL=http://your-server:port/v1/chat/completions
VISION_MODEL=Qwen3-VL-32B
VISION_API_KEY=your-api-key在 Claude Code 的 ~/.mcp.json 中添加：
{
"mcpServers": {
"vision": {
"command": "node",
"args": ["/path/to/vision-mcp-server/dist/index.js"],
"env": {
"VISION_BASE_URL": "http://your-server:port/v1/chat/completions",
"VISION_MODEL": "Qwen3-VL-32B",
"VISION_API_KEY": "your-api-key"
}
}
}
}将 /path/to/vision-mcp-server 替换为实际安装路径。
Unclaimed servers have limited discoverability.
If you are the server author, to access and configure the admin panel.
OCR, transcription, file extraction, and image generation for AI agents via MCP.
Generate images, GIFs, and PDFs from HTML, URLs, or templates — from your AI agent.
Generate, edit and upscale AI video and images from any agent via VicSee.
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/Loveacup/vision-mcp-server'
If you have feedback or need assistance with the MCP directory API, please join our Discord server