@pipeworx/rescuegroups

Adoptable animals listed by rescues and shelters on RescueGroups.org, and the organizations themselves — name, breed, age, sex, size, temperament flags, photos, adoption fee, and the rescue's contact details, adoption process and areas served. Searchable by species, postal-code radius, breed and trait.

Part of Pipeworx — an MCP gateway connecting AI agents to 1576+ live data sources.

Tools

• rescuegroups_adoptable(species?, postalcode?, lat?, lon?, miles?, kilometers?, breed?, sex?, age_group?, size_group?, name?, good_with_kids?, good_with_dogs?, good_with_cats?, views?, org_id?, include_adopted?, sort?, limit?, page?, _apiKey) — the main search. Answers "adoptable beagles within 50 miles of 15206", "cats good with kids in Pittsburgh", "dogs needing a foster". Returns the animal's full attribute set with breeds, pictures, species and listing organization resolved inline.

• rescuegroups_animal(animal_id, _apiKey) — one animal in full, with every picture, colour, pattern, status and the listing org.

• rescuegroups_organizations(postalcode?, lat?, lon?, miles?, kilometers?, name?, state?, city?, type?, org_id?, sort?, limit?, page?, _apiKey) — rescues and shelters. Answers "shelters within 25 miles of me", "who is RescueGroups org 1234".

• rescuegroups_species(limit?, page?, _apiKey) — the species vocabulary, which is what rescuegroups_adoptable's species argument expects.

• rescuegroups_breeds(species?, limit?, page?, _apiKey) — breeds per species, for building an exact breed filter.

Related MCP server: mcp-dogsapi

Auth

BYO-key only. Pipeworx holds no RescueGroups key and this pack never reads a platform one — _apiKey is required in every tool's input schema, so the gateway's disclosure surface tells a caller before the call, not after a refusal.

Request a free public key at https://rescuegroups.org/services/request-an-api-key/. It is a short form (name, email, organisation, site URL, which species and whether you want a national search). It is open to individual developers — "not incorporated" is an accepted incorporation status, and the API docs state plainly that you do not need a RescueGroups account, only a public key. Staff review the form and email the key back, so it is not instant self-serve; the refusal message says so rather than promising a key in one click.

Note for anyone reading docs/paid-upgrade-watchlist.md: its entry filing RescueGroups under "signup not obtainable by a normal dev — requires being an animal rescue with a vet, shelter, euthanasia policy" describes the partner onboarding at /partner-requirements-and-expectations/, which is how a rescue gets its animals listed. The public API key form is a different, open one. That entry has been corrected.

Data source

• https://api.rescuegroups.org/v5 — RescueGroups.org v5 public API, JSON:API over HTTPS. Full reference (unauthenticated, readable without a key): https://api.rescuegroups.org/v5/public/docs

Things the next person would otherwise rediscover the hard way

An invalid key is a silent zero, not an error. Probed live 2026-09-05:

GET /v5/public/animals/search/available/dogs/?limit=1     (no Authorization)
  -> 401 application/vnd.api+json
     {"errors":[{"status":401,"source":{"header":"Authorization"},
       "title":"Missing authorization header",
       "detail":"Missing authorization header - apikey authentication required"}]}

GET /v5/public/animals/search/available/dogs/?limit=1     (Authorization: <junk>)
  -> 200 OK, Content-Length: 0, Content-Type: text/plain;charset=UTF-8

A rejected key returns HTTP 200 with a completely empty body, not a 401 — and it does so whether or not the required Content-Type header is present (checked both ways; the empty 200 is not an artefact of omitting it). Passed through naively that reads as "there are no adoptable dogs near you", which is both wrong and unfixable by the user because nothing points at their key. This pack therefore treats any empty non-JSON 200 as an auth failure and says so. See docs/silent-zero-policy.md.

Consequently the three states a caller can be in read differently, on purpose:

The key never leaves the Authorization header. It is not put in a URL (a URL is logged by every hop whether or not the call succeeds), it is deleted from args before any handler runs so it cannot reach a usage record, and every string relayed from upstream is passed through a redactor first — so even if RescueGroups started quoting the key back in an error detail, the caller-facing message would not carry it. Verified by a test that plants a recognisable secret and asserts it appears in none of the URL, body, result or error paths.

Keys are scoped at issue time. The request form asks which species and whether you want just your own organization or a national search. A key issued narrowly returns fewer rows — or none — for a query outside its scope, which looks identical to a genuinely empty result. The rejected-key message names this possibility so a caller with a legitimately narrow key is not sent chasing a bad-key theory.

Content-Type: application/vnd.api+json is required on every request, including GETs, per the docs. Sent unconditionally here.

Views stack in the URL path, filters go in a POST body. available, species (dogs), and traits (haspic, urgent, isneedingfoster) are "views" and concatenate: /public/animals/search/available/dogs/haspic/. Anything else — breed, sex, size, postal-code radius — is a filters / filterRadius object POSTed to the same search URL. The pack picks GET or POST based on whether the caller supplied anything needing a body.

Distance searches need a location and a radius. Half of one is rejected locally rather than sent, because RescueGroups silently returns unfiltered results for a location with no radius, which looks like the radius was honoured.

Org search with no type needs two calls. rescue and shelter are the only org views, a filtered org search requires a view in the path, and the two cannot be stacked (they would AND to nothing). With no type the pack queries both and merges, and says so in a note — pages is not meaningful on that path, so pass type if you need to page.

Null attributes are omitted, not nulled (v5 change log, 2019-10-03), and the field set grows over time. The response mapper is deliberately generic — it flattens JSON:API attributes wholesale and resolves relationships against included one level deep — rather than listing fields, so new upstream fields appear instead of being silently dropped. raw always carries the untouched body.

Rate limits are unpublished. The docs say only that the API returns 429 for "an abnormal number of requests" and asks callers to cache reference data. Cache rescuegroups_species and rescuegroups_breeds; they change rarely.

There is a test environment at test1-api.rescuegroups.org with the same routes, but it also requires a key, and its data may be weeks stale and may not be used in production. Not wired up here.

Quick Start

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

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

What this endpoint actually serves

tools/list at https://gateway.pipeworx.io/rescuegroups/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 1576+ 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

This pack takes your own API key (_apiKey) — we don't front one for it, so there's no curl here that would run without it. Inspect any tool: GET https://gateway.pipeworx.io/v1/tools/rescuegroups_adoptable. 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": {
    "rescuegroups": {
      "command": "npx",
      "args": ["-y", "@pipeworx/mcp-rescuegroups"]
    }
  }
}

Or run it directly to confirm it starts:

npx -y @pipeworx/mcp-rescuegroups

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 Rescuegroups data" })

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

More

• Docs and guides

• pipeworx.io

License

MIT