jenkinsm13/metashape-mcp

View on GitHub ↗

AI-powered photogrammetry automation — MCP server for Agisoft Metashape Professional 2.3+. Control the full pipeline (alignment, dense cloud, mesh, texture, DEM, orthomosaic, export) using natural language through Claude or any MCP-compatible AI assistant.

31 ★5 forksPythonUpdated 4mo ago

What you need to know

MCP server bringing AI-powered automation to Agisoft Metashape Professional with 106 tools, 10 resources, and 6 prompts covering the full photogrammetry pipeline.

Install

Requires Agisoft Metashape Professional 2.3+
Clone repo and start the server inside Metashape via scripts/start_mcp_server.py
Dependencies (mcp, fastmcp) auto-installed on first run
Headless option: metashape -platform offscreen -r start_mcp_headless.py

Usage

  • Start the HTTP server in Metashape (default http://127.0.0.1:8765/mcp)
  • For Claude Code, use the stdio proxy (python -m metashape_mcp.proxy) to avoid timeouts
  • Ask the agent to process drone surveys, build 3D models, generate DEMs/orthomosaics, export deliverables

Key features

  • 106 tools across 15 modules covering the photogrammetry pipeline
  • 10 resources for real-time project state inspection, 6 guided workflow prompts
  • Streamable HTTP transport plus stdio proxy with no timeout
  • Non-blocking GUI and headless mode for remote/CI use
  • execute_python tool for arbitrary code inside Metashape
  • Progress reporting for long operations
  • MIT license

Best for

AI-driven photogrammetry, drone mapping, and 3D reconstruction in Metashape.

Clients: Claude Code · Claude Desktop

Reviewed 2026-08-11

Topics

3d-reconstructionagisoftclaudedrone-mappingmcpmetashapemodel-context-protocolorthomosaicphotogrammetrypoint-cloud
Stars
31★
Forks
5
Language
Python
License
MIT
Created
2026-02-21
Last push
2026-05-07