mcpField
O
Overture Places

Overture Places

Remote
pipeworx.ioMaps & Travel0Updated 3w ago

About Overture Places

Overture Places is an MCP server for searching business and point-of-interest data in the US and UK. It supports place discovery, company location lookups by domain, and area counts, with confidence filtering and source attribution.

Key features

  1. 01Search places by name, category, location, or city
  2. 02Find company locations by website domain
  3. 03Count places and categories in an area
  4. 04Filter by confidence and closed status
  5. 05Return source attribution and data release

Use cases

  1. 01Find nearby businesses matching a name or category
  2. 02Map a company’s locations from its website domain
  3. 03Count points of interest within a geographic area

MCP support

Exposes place search, domain lookup, and area-count tools to MCP clients. It can run as a local stdio server or through the Pipeworx gateway.

Configuration

Remote servers need no local install. Pick your client, then copy the config or install in one click.

Endpointhttps://gateway.pipeworx.io/overture-places/mcp
Copy install command
Claude CodeTerminal
Cursor.cursor/mcp.json
VS Code.vscode/mcp.json
Claude Desktopclaude_desktop_config.json
Windsurf~/.codeium/windsurf/mcp_config.json

Run commands in a terminal; paste JSON into the client’s MCP config and restart. Servers that need sign-in will prompt on first connect.

Tools

Details

From the project README

@pipeworx/overture-places

Business and point-of-interest locations in the United States and the United Kingdom, from the Overture Maps Foundation places theme. Each place carries a name, category, existence confidence score, street address, websites, phones, emails, social links, brand and contributing source record ids. Use it to find places near a point or in a city, to map every location of a company by its website domain, and to count what is in an area.

Part of Pipeworx — an MCP gateway connecting AI agents to 1764+ live data sources. This is an independent, unofficial integration — not affiliated with, endorsed by, or published by the upstream provider.

Tools

  • places_search(name?, category?, near{lat,lon} + radius_m | bbox | city + country [+ region], min_confidence?, include_closed?, limit?): businesses matching a name and/or category. Results are ranked by match quality and then by distance (for near) or confidence. Example: the Starbucks branches within 3 km of downtown Austin.
  • places_by_domain(domain, country?, region?, min_confidence?, include_closed?, limit?): every location whose website is on a company's domain, which gives that company's physical footprint. Subdomains and paths fold to the registrable domain (https://www.starbucks.com/store-locator → starbucks.com). Returns the total plus counts by country and by state/region.
  • places_in_area(near + radius_m | bbox | city + country, category?, min_confidence?, include_closed?, limit?): the count of places in an area, a breakdown by category, and the highest-confidence places.

Every response starts with attribution, a source block (licence, release, terms URL) and data_as_of, which is the Overture release the answer came from (for example 2026-09-23.1).

Confidence

Overture gives every place an existence confidence from 0 to 1 and calls it "the primary tool for filtering potential junk data". By default this pack hides places below 0.7, and places marked permanently_closed. In the measured sample the distribution breaks sharply at 0.7, and the rows below it were mostly junk. The default keeps about 68% of places. Every place shows its confidence, and every response states the threshold it applied and how many places it hid (hidden_by_filters). To see everything, pass min_confidence: 0 and include_closed: true.

Coverage and limits

  • Countries: US and GB. An area or city outside them returns found:false, reason:"outside_coverage" and lists the covered countries.
  • One call scans at most 64 data blocks, nearest the centre of the area first. That is about 250,000 places. For a larger area (a whole metro searched by name), the response sets complete:false and explains in coverage_warning what was not searched. places_in_area still gives exact or bounded totals from its index. Narrow the area with near + radius_m for a complete answer.
  • City lookup resolves a locality name from the addresses in the data. The region field is consistent in the US (state codes; full state names are accepted too) but inconsistent in the UK, so leave region out for UK cities. A name that matches several places (several Springfields) uses the largest and lists the others in other_places_with_this_name.
  • Platform domains are not indexed. Social profiles, link pages, site builders and listing sites (instagram.com, linktr.ee, wixsite.com, yelp.com and similar) identify no single company. places_by_domain refuses them with reason:"platform_domain".
  • Very large domains: the 5,000 highest-confidence locations per domain are indexed. The response always gives the true total in locations_listing_domain and says when the cap applied.
  • Overture warns that the theme "is known to contain duplicates, a high junk rate, and low property completeness". Confidence filtering helps with junk but does not remove duplicates.

Auth

Keyless. The pack answers only through the Pipeworx gateway (https://gateway.pipeworx.io/mcp). A standalone install has no Overture release to read, so every call fails with an error that says so.

Data sources

  • Overture Maps Foundation places theme, monthly GeoParquet releases: https://docs.overturemaps.org/guides/places/, at s3://overturemaps-us-west-2/release/<release>/theme=places/type=place/. The overture-places-ingest workflow checks for a new release every week (design: docs/overture-places-design.md).
Licence and attribution

The places theme is published under CDLA-Permissive-2.0 and Apache-2.0, depending on the source. Redistribution is permitted, including commercial use, with attribution. Overture states that the theme contains no OpenStreetMap data, so ODbL does not apply. Attribution, as Overture publishes it (https://docs.overturemaps.org/attribution/):

© Overture Maps Foundation (https://overturemaps.org), places theme. Data from Meta. Available under CDLA Permissive 2.0. Data from Microsoft. Available under CDLA Permissive 2.0. Data from PinMeTo. Available under CDLA Permissive 2.0. Data from Krick. Available under CDLA Permissive 2.0. Data from RenderSEO. Available under CDLA Permissive 2.0. Data from DAC. Available under CDLA Permissive 2.0. Data from BrightQuery. Available under CDLA Permissive 2.0. Data from Foursquare. Copyright 2024 Foursquare Labs, Inc. All rights reserved. Available under Apache 2.0. Foursquare data was transformed to the Overture schema. Data from AllThePlaces. Available under CC0 1.0.

Each response's attribution names the Overture release and the per-source credit lines for the sources in that response. Each place lists its own contributing datasets in sources and their licences in source_licenses.

Quick Start

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

{
  "mcpServers": {
    "overture-places": {
      "url": "https://gateway.pipeworx.io/overture-places/mcp"
    }
  }
}
What this endpoint actually serves

tools/list at https://gateway.pipeworx.io/overture-places/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 1764+ 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/places_search \
  -H 'Content-Type: application/json' \
  -d '{"name":"Starbucks","near":{"lat":30.2672,"lon":-97.7431},"radius_m":3000,"limit":5}'

No account needed for the first calls. Inspect any tool: GET https://gateway.pipeworx.io/v1/tools/places_search. 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": {
    "overture-places": {
      "command": "npx",
      "args": ["-y", "@pipeworx/mcp-overture-places"]
    }
  }
}

Or run it directly to confirm it starts:

npx -y @pipeworx/mcp-overture-places

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 Overture Places data" })

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

More

License

MIT

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.