Mobility Database MCP Server

@pipeworx/mobility-database

The open catalogue of public-transport data feeds maintained by MobilityData — ~4,500 GTFS Schedule feeds and ~1,900 GTFS-Realtime feeds across ~70 countries, each with the operator's own download URL, licence and bounding box. The successor to TransitFeeds.

Part of Pipeworx — an MCP gateway connecting AI agents to 1683+ live data sources.

Tools

• mobilitydb_search_feeds(query?, country?, municipality?, provider?, data_type?, include_inactive?, limit?) — find the transit feed for an agency, city or country.
• mobilitydb_gtfs_rt_feeds(country?, query?, entity_type?, group_by_agency?, limit?) — realtime feeds (vehicle positions / trip updates / service alerts) with the schedule feed each is keyed against.
• mobilitydb_feed(feed_id) — everything the catalogue records about one feed, plus its sibling realtime streams and any superseding entry.

Auth

Keyless.

MobilityData also runs an authenticated REST API at api.mobilitydatabase.org, which takes a bearer token minted from a refresh token issued on free signup at https://mobilitydatabase.org/sign-up. This pack deliberately does not use it and declares no platform key: the catalogue export carries the same feed inventory with no credential, and declaring a key env var the gateway does not hold un-sinks routing toward a tool that could only refuse. What the authenticated API would add is per-version dataset history and MobilityData's own validation reports; if that is wanted, PLATFORM_MOBILITYDB_TOKEN and the code that uses it land together in one change.

Data sources

• https://files.mobilitydatabase.org/feeds_v2.csv — the catalogue export (~6,475 rows including superseded entries).

Traps, verified live 2026-09-17

• It is real CSV, not line-splittable. Feed names and notes contain commas, quotes, CJK text and nested parentheses — "TC_大有巴士(公總) (Dayou bus (public bus))". Splitting on , puts an operator's contact email in the data_type column. There is an RFC 4180 parser in src/index.ts for exactly this.
• One agency publishes three realtime rows, one per entity_type: vp vehicle positions, tu trip updates, sa service alerts, each with a different URL. Counting "realtime feeds" without collapsing on static_reference triples the agency count.
• A realtime feed is unusable without its schedule feed. static_reference names it; realtime trip ids only resolve against that specific GTFS feed.
• Some feeds need the CONSUMER's own credential at the transit agency (urls.authentication_type 1 or 2, parameter named in urls.api_key_parameter_name). That credential is between the caller and the agency; it is reported so a caller knows before fetching.
• redirect.id marks a superseded entry that is still present in the export and would otherwise look like a live feed.

Quick Start

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

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

What this endpoint actually serves

tools/list at https://gateway.pipeworx.io/mobility-database/mcp returns the tools in the table above plus the shared Pipeworx meta-tools — ask_pipeworx, discover_tools, 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 1683+ 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/mobilitydb_search_feeds \
  -H 'Content-Type: application/json' \
  -d '{"query":"BART","country":"US","limit":2}'

No account needed for the first calls. Inspect any tool: GET https://gateway.pipeworx.io/v1/tools/mobilitydb_search_feeds. Find one: POST https://gateway.pipeworx.io/v1/tools/search_packs with {"query":"..."}.

Standalone (no gateway account)

This package also runs as a local stdio MCP server — no Pipeworx account, no gateway round-trip:

{
  "mcpServers": {
    "mobility-database": {
      "command": "npx",
      "args": ["-y", "@pipeworx/mcp-mobility-database"]
    }
  }
}

Or run it directly to confirm it starts:

npx -y @pipeworx/mcp-mobility-database

It speaks MCP over stdin/stdout and answers initialize/tools/list/tools/call for only this pack's tools — none of the shared meta-tools the gateway connection above adds. Same source, same tools, no ask_pipeworx routing.

Using with ask_pipeworx

Instead of calling tools directly, you can ask questions in plain English — this works on the pack endpoint above as well as on the full gateway:

ask_pipeworx({ question: "your question about Mobility Database data" })

The gateway picks the right tool and fills the arguments automatically.

More

• Docs and guides
• pipeworx.io

License

MIT