An MCP server that provides AI agents with geospatial analysis capabilities using GDAL/Rasterio. Includes a reflection system that requires agents to justify their methodological decisions before executing operations.

Installation

uvx (Recommended)

uvx --from gdal-mcp gdal --transport stdio

Testing

uv run pytest test/ -v

# With coverage
uv run pytest test/ --cov=src --cov-report=term-missing

Configuration

Claude Desktop

{
  "mcpServers": {
    "gdal-mcp": {
      "command": "uvx",
      "args": ["--from", "gdal-mcp", "gdal", "--transport", "stdio"],
      "env": {
        "GDAL_MCP_WORKSPACES": "/path/to/your/geospatial/data"
      }
    }
  }
}

Available Tools

Tool	Description
raster_info	Check metadata (CRS, resolution, bands, nodata)
raster_convert	Convert formats with compression and overviews (COG support)
raster_reproject	Transform CRS (with reflection)
raster_stats	Statistical analysis with histograms
vector_info	Check metadata (CRS, geometry, attributes)
vector_reproject	Transform CRS (with reflection)
vector_convert	Format migration (SHP ↔ GPKG ↔ GeoJSON)
vector_clip	Spatial subsetting
vector_buffer	Proximity analysis
vector_simplify	Geometry simplification
store_justification	Cache epistemic reasoning (used internally)

Capabilities

• Reflection middleware with pre-justification for CRS selection and resampling methods
• Persistent cache with 75% hit rate in multi-operation workflows
• Cross-domain cache sharing — CRS justification works for BOTH raster and vector data
• Comprehensive toolset for raster and vector data
• Full type safety with mypy strict mode
• Workspace security with path validation middleware
• Python-native stack using Rasterio/PyProj/pyogrio
• Real-time feedback via FastMCP Context API
• Workspace catalog for autonomous file discovery
• Intelligent metadata for format determination

Environment Variables

Required

• GDAL_MCP_WORKSPACES - Path to directory with geospatial data

Usage Examples

Reproject this DEM to Web Mercator for my web map

I need to reproject this DEM to UTM for accurate slope analysis, then reproject this vector layer to the same CRS for overlay

Notes

Built with a reflection system that requires AI agents to justify methodological decisions before execution, creating an audit trail for reproducible geospatial science. Includes 72 passing tests and comprehensive documentation, including Quick Start, Tools Reference, Vision, and Changelog.