origin-mcp

简体中文

origin-mcp is a local Model Context Protocol (MCP) server that lets AI assistants control Origin/OriginPro on Windows. It connects through OriginLab's Python automation interface and exposes tools for importing data, editing worksheets, creating and refining graphs, running Origin analyses, exporting figures, and managing the Origin application lifecycle.

This project is still in a testing stage. Trying it on real Origin workflows, reporting issues, suggesting improvements, and opening pull requests are all welcome.

Highlights

• Import CSV, TSV, TXT, DAT, XLS, and Excel data into Origin worksheets.
• Read, write, sort, clear, and export worksheet data.
• Create and refine common 2D, 3D, contour, statistical, and specialized plots.
• Run Origin analyses such as fitting, smoothing, integration, peak finding, and descriptive statistics.
• Export figures and projects through a local Origin GUI bridge.

Requirements

• Windows
• Origin or OriginPro installed and licensed
• Origin/OriginPro 2026 is the primary tested target; other Origin versions are not currently guaranteed
• Origin's originpro package and pywin32 (for the Origin-side bridge)

Python version support

origin-mcp runs as two cooperating processes, and the supported Python versions differ by role:

• MCP server core (the python -m origin_mcp process, which only talks to the bridge over localhost): Python 3.10+. Local checks currently pass on Python 3.12 and 3.14, and 3.10/3.11/3.13 are expected to work as well.
• Origin bridge (addon.py): runs inside Origin's own embedded Python, so its version is whatever your Origin install ships — there is nothing to pick.
• External originpro automation (only if you import originpro/pywin32 directly into the MCP core process instead of using the bridge): use Python 3.11 or 3.12. Python 3.13 is experimental and 3.14+ is not a supported external target, because Origin's automation packages lag newer Python releases — start the bridge inside Origin's embedded Python instead.

Agentic Setup

Copy this to your AI agent and let it self-configure:

Fetch and follow this bootstrap guide end to end:
https://raw.githubusercontent.com/Ge-Shun/origin-mcp/main/docs/agentic/origin-mcp-bootstrap.md

MCP Configuration

Example MCP client configuration:

{
  "mcpServers": {
    "origin": {
      "command": "python",
      "args": ["-m", "origin_mcp"]
    }
  }
}

If python is not the Python 3.10+ interpreter you installed origin-mcp into, use that interpreter's absolute python.exe path instead. More examples are in docs/mcp-config.md.

Start the Origin Bridge

The bridge runs inside Origin's own Python so originpro stays on Origin's UI thread. There is nothing to configure — start it once per Origin session:

1. Open Origin, then open its Python Console.
2. Paste this single line (replace the path with your checkout):

import runpy; runpy.run_path(r"C:\path\to\origin-mcp\addon.py", run_name="__main__")

A Bridge is running inside Origin. box confirms startup. Keep that console running while you use the tools.

To stop, just ask your MCP assistant to shut the Origin bridge down — it calls origin_bridge_shutdown, so no extra terminal or console input is needed. If you are not using an assistant, double-click scripts\stop-bridge.cmd (or run python scripts\stop_bridge.py) to send the same shutdown request. The serving console returns to its prompt and Origin stays open.

If a package is missing or the bridge will not start, see docs/origin-bridge.md.

License

MIT. See LICENSE.