Coming Soon: Igor Pro MCP Server

Connect an AI agent to Igor Pro using our official Model Context Protocol (MCP) server. MCP provides a standardized way for large language models (LLMs) to communicate with external applications. Most modern LLMs can be used with this system (such as Claude, GPT, and Gemini) as long as the application hosting the chat supports MCP tools, such as Visual Studio Code, Claude Code, Cursor, and Codex.

The Igor Pro MCP server gives AI agents a powerful set of tools for working directly with Igor. Among other things, AI agents will be able to:

• Compile and execute Igor code, receiving compilation errors as feedback
• Directly call built-in and user-defined functions
• Search the official documentation
• Capture images of Igor windows for visual feedback
• Inspect the state of the application, and retrieve information about the current experiment
• Launch and quit Igor

These tools let agents carry out tasks, check results, and iterate on their work using feedback from Igor.

See it in action

Here, we prompt the AI agent:

I want to visualize the crystal structure of green fluorescent protein (GFP). Can you download the structure from the protein databank and make a 3D plot with it?

The agent writes Igor code to import the downloaded coordinates and produces a 3D interactive graphic. The agent gets visual feedback from Igor along the way to iterate and improve the final result. Note, this video is sped up, and covers about 4 minutes in real time.