MCP Read Images

MCP Read Images is an MCP server that provides vision-based image analysis through the OpenRouter API. It connects MCP-compatible development environments and assistants directly to advanced multimodal vision models, including Claude-3.5-sonnet and Claude-3-opus. Software engineers, designers, and AI developers use this server to inspect visual assets, analyze diagrams, and extract visual context from local files without leaving their coding workflow. The server exposes an analyze_image tool that handles automatic image resizing, JPEG format conversion, and quality optimization before forwarding queries to the chosen vision model. Users can ask open-ended questions about visual contents or request specific details about local image files. MCP Read Images supports flexible configuration, letting users establish default models across their setup via environment variables or selectively override the vision model per tool request for fine-tuned multimodal performance.

Category: AI & LLM Tooling

Tags: image-analysis, multimodal, openrouter, vision

Visit MCP Read Images

How to install and configure MCP Read Images

1. Install the server package globally or in your project using npm: bash npm install @catalystneuro/mcp_read_images Alternatively, clone and build from the repository: bash git clone https://github.com/catalystneuro/mcp_read_images.git cd mcp_read_images npm install npm run build 2. Open your MCP configuration file (such as cline_mcp_settings.json for Cline in VSCode). 3. Add the read_images configuration under mcpServers with your OpenRouter API key: json { "mcpServers": { "read_images": { "command": "read_images", "env": { "OPENROUTER_API_KEY": "your-api-key-here", "OPENROUTER_MODEL": "anthropic/claude-3.5-sonnet" }, "disabled": false, "autoApprove": [] } } } 4. Restart your MCP client to activate the server.

What you can do with MCP Read Images

• Inspect local UI screenshots directly from an editor to detect visual discrepancies, alignment flaws, or missing design elements. * Query architecture diagrams or flowcharts to convert visual components and relationships into structured implementation plans. * Extract text, tables, and data values from locally saved scanned documents and receipts via vision models. * Verify exported chart visualizations, generated assets, and test output images automatically within automated agent loops.

Key facts

• https://github.com/catalystneuro/mcp_read_images
• AI & LLM Tooling, Design, Media & Creative
• image-analysis, multimodal, openrouter, vision

Part of MCP Servers

What is MCP Read Images?

MCP Read Images is a Model Context Protocol server that enables AI clients to inspect and analyze local images using OpenRouter vision models. It automatically optimizes and resizes local image files before passing them along with questions to multimodal models like Claude 3.5 Sonnet.

How do I install MCP Read Images?

You can install the package by running npm install @catalystneuro/mcp_read_images. Alternatively, clone the GitHub repository, run npm install, and build with npm run build, then register the read_images command in your MCP client settings file alongside your OpenRouter API key.

What models does MCP Read Images support?

The server connects via OpenRouter and has been tested with anthropic/claude-3.5-sonnet and anthropic/claude-3-opus-20240229. You can define a model globally in the environment settings or specify a different model per individual tool call.

Is MCP Read Images open source?

Yes, MCP Read Images is open-source software released under the MIT License. The complete source code and licensing details are available on GitHub in the catalystneuro/mcp_read_images repository.