goes-mcp

MCP server for NOAA GOES satellite imagery — visible, infrared, water vapor, and composite products from GOES-18 (West) and GOES-19 (East).

Features

• 16 ABI spectral bands — visible, near-IR, and infrared channels
• 6 composite products — GeoColor, AirMass, Sandwich, Fire Temperature, Dust, Derived Motion Winds
• Multiple coverages — CONUS, Full Disk, and 5 regional sectors (SE, NE, Caribbean, Tropical Atlantic, Puerto Rico)
• Image embedding — returns JPEG images directly via MCP Image type
• Flexible output — image (default), markdown with URL, or JSON metadata

Data Sources

• NOAA STAR CDN (cdn.star.nesdis.noaa.gov) — pre-rendered JPEG imagery at multiple resolutions
• RAMMB/CIRA SLIDER (slider.cira.colostate.edu) — timestamp discovery for historical imagery

No API keys required. All data is freely available from NOAA.

Tools

Installation

# With uv (recommended)
uvx goes-mcp

# With pip
pip install goes-mcp

Usage with Claude Desktop

Add to your Claude Desktop config:

{
  "mcpServers": {
    "goes": {
      "command": "uvx",
      "args": ["goes-mcp"]
    }
  }
}

Example Queries

• "Show me the latest GOES-East GeoColor image of the Continental US"
• "Get a satellite view of the Caribbean right now"
• "What satellite imagery is currently available?"
• "Show me infrared band 13 for the full disk view"
• "Get the latest fire temperature composite for the Southeast US"

Development

cd servers/goes-mcp
uv sync --group dev

# Run unit tests
uv run pytest tests/ --ignore=tests/test_live.py --ignore=tests/test_mcp_protocol.py -v

# Run MCP protocol test
uv run pytest tests/test_mcp_protocol.py -v

# Run live integration tests (requires network)
uv run pytest tests/test_live.py -v

# Lint
uv run ruff check src/ tests/ --select E,F,W --ignore E501
uv run ruff format --check src/ tests/

License

MIT