gis-rs-mcp

gis-rs-mcp is an enhanced GIS MCP server build with lightweight remote-sensing tools.

It keeps the Python module name gis_mcp and the command name gis-mcp, while publishing under a separate PyPI distribution name because the original gis-mcp project name is already owned on PyPI.

Install

pip install gis-rs-mcp

Optional data/visualization extras:

pip install "gis-rs-mcp[satellite-imagery,land-cover,visualize]"

Run the MCP server:

gis-mcp

Remote-Sensing Tools

Built-in sensor presets:

• Sentinel-2
• Landsat 4/5/7
• Landsat 8
• Landsat 9
• MODIS
• NAIP
• PlanetScope-style stacks

Common indices:

• NDVI
• GNDVI
• NDWI
• MNDWI
• NDMI
• NDBI
• NBR
• NBR2
• EVI
• SAVI
• MSAVI
• NDRE

Useful tools include:

• list_remote_sensing_sensors
• get_sensor_band_preset
• list_spectral_indices
• compute_spectral_index
• get_satellite_download_preset
• describe_remote_sensing_data_sources

compute_spectral_index works on local or HTTPS GeoTIFF rasters and can use either a sensor preset or an explicit band map.

Data Accounts

The default STAC download path is Microsoft Planetary Computer, which usually does not require a personal account for Sentinel-2, Landsat, MODIS, NAIP, and ESA WorldCover. Other providers are optional and commonly require credentials:

• Copernicus Data Space: Copernicus account/token
• USGS EarthExplorer/M2M: USGS EROS account
• NASA Earthdata/LAADS: NASA Earthdata Login token
• Sentinel Hub: Sentinel Hub account/quota
• Google Earth Engine: Google account/project authentication

The MCP itself does not require a VPN; network reliability depends on the user's route to the selected data service.