Planetary Computer MCP server

@pipeworx/planetary-computer

Earth-observation imagery catalogue from Microsoft's Planetary Computer — search its open satellite and environmental collections (Sentinel, Landsat, NAIP, Daymet, ERA5, biodiversity) by area and date and get item ids, footprints, dates and asset URLs.

Part of Pipeworx — an MCP gateway connecting AI agents to 1679+ live data sources.

Tools

the catalogue with licence, extent and date range. Call this first: search refuses a query that names no collection.

• planetarycomputercollections(contains?, prefix?, limit?) — the datasets in

including the asset keys (bands) its items carry.

• planetarycomputercollection(collection_id) — one dataset in full,

— which scenes cover an area on given dates, with cloud cover and per-band asset URLs.

• planetarycomputersearch(collections, bbox?, datetime?, ids?, limit?, max_assets?)

Auth

Keyless for metadata. Asset BYTES are a separate step: most collections require a short-lived read token from https://planetarycomputer.microsoft.com/api/sas/v1/token/{collection} (itself keyless, ~1 hour) appended to the href. The href alone is not a working download link and reads like a dead one if you try it cold.

Data sources

• — STAC API 1.0.

Shared client: shared/src/stac.ts (stacCollections / stacCollection / stacSearch), also used by earth-search, copernicus-dataspace and meteoswiss. Fix paging or an error message there and all four improve.

Traps

required — a bare string naming no fix. The shared client converts it into a message that names planetarycomputer_collections.

• POST /search with no collections answers **HTTP 422 collection is

sentinel-2-c1-l2a on Earth Search. Passing a sibling catalogue's id returns an empty result, not an error.

• Collection ids are per-catalogue. sentinel-2-l2a here is

truncated when it stops.

• GET /collections pages; the shared client follows rel="next" and reports

Quick Start

Add to your MCP client (Claude Desktop, Cursor, Windsurf, etc.):

{
  "mcpServers": {
    "planetary-computer": {
      "url": "https://gateway.pipeworx.io/planetary-computer/mcp"
    }
  }
}

What this endpoint actually serves

tools/list at https://gateway.pipeworx.io/planetary-computer/mcp returns the tools in the table above plus the shared Pipeworx meta-tools — askpipeworx, discovertools, search_within, remember/recall and the rest of the gateway-wide set. So the tool count you see is larger than this table: a single-pack endpoint currently lists roughly 30 shared tools alongside the pack's own. The connection's initialize response states its exact scope, and is the authoritative answer for a given day.

This is deliberate, not multiplexing by accident. The meta-tools are what let a scoped connection answer a question this pack does not cover — via ask_pipeworx, which routes across the whole catalog — without you adding a second MCP server. There is currently no way to mount a pack endpoint without them; if the extra schemas cost you more context than the routing is worth, connect to the full gateway once rather than to several pack endpoints.

Or connect to the full Pipeworx gateway to get every pack's tools listed directly, instead of just this one's:

{
  "mcpServers": {
    "pipeworx": {
      "url": "https://gateway.pipeworx.io/mcp"
    }
  }
}

Both URLs reach the same gateway and the same 1679+ data sources. The only difference is which pack's tools are listed directly; ask_pipeworx reaches all of them from either one.

No MCP client? Call it over HTTP

curl -X POST https://gateway.pipeworx.io/v1/tools/planetary_computer_collections \
  -H 'Content-Type: application/json' \
  -d '{"contains":"sentinel","limit":5}'

Shortened. The full README is on GitHub.