Inaturalist

What Inaturalist does

Inaturalist exposes read-only access to iNaturalist citizen-science data through MCP. It covers observations of any organism, including plants, fungi, insects, birds, mammals, reptiles, and amphibians. The integration is independent and unofficial; it is not affiliated with or endorsed by iNaturalist.

Three domain tools form the main interface:

• search_observations finds sightings using optional taxon, place, year, threatened-status, quality-grade, and page-size filters. Results include photographed observations with coordinates, dates, and observer information.
• search_taxa looks up common or scientific names and returns taxonomic identifiers, rank, ancestry, conservation status, and a photo URL.
• top_species ranks the most frequently observed species for a place, optionally over a year or date range, and includes observation counts and taxonomic details.

How it works

The hosted MCP endpoint forwards requests to the public iNaturalist REST API at https://api.inaturalist.org/v1/. Read-only requests require no iNaturalist key or account. Taxon and place names can be resolved automatically, while a taxon ID returned by search_taxa can help disambiguate similar names or support later filtering.

Connect an MCP client to https://gateway.pipeworx.io/inaturalist/mcp to use the hosted version. The same gateway also provides a full endpoint at https://gateway.pipeworx.io/mcp; the pack endpoint lists the Inaturalist tools directly, while the full endpoint lists tools from the wider Pipeworx catalog. The pack endpoint also exposes shared Pipeworx meta-tools, including ask_pipeworx, discover_tools, search_within, and memory tools. ask_pipeworx can route a natural-language request across the gateway’s catalog.

The service can also be called without an MCP client through HTTP. Tool requests use endpoints such as /v1/tools/search_observations, and tool metadata can be inspected with a GET request to the corresponding tool path.

Setup and configuration

For a hosted connection, add the MCP URL to the client configuration:

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

A local npm bridge starts with npx and forwards requests to the same hosted gateway:

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

The npm bridge needs network access and has no offline mode. Anonymous use is limited to 50 calls per day. A free Pipeworx account and PIPEWORX_API_KEY raise the allowance to 200 calls per day; the key is sent as a Bearer token. Set PIPEWORX_META_TOOLS=1 if the local bridge should also list the shared gateway tools. When using an HTTPS proxy, Node’s fetch behavior may require NODE_USE_ENV_PROXY=1, depending on the Node version.

Tools and capabilities

search_observations fits questions such as sightings of a species in a region during a particular year. It accepts common or scientific taxon names, place names, year, threatened status, quality grade, and result count. search_taxa is useful before observation searches when a name is ambiguous. top_species resolves a place name to a place identifier and returns ranked species for that location.

Observation results can be paired with GBIF data for cross-validation, although GBIF is not provided by these three tools. The service is read-only, so it is intended for querying and analysis rather than submitting or editing observations.

Limitations and notes

The integration depends on the hosted Pipeworx gateway and the public iNaturalist API. Local use therefore still requires internet access. Anonymous npm-bridge calls have a daily limit, while authenticated calls are metered through the Pipeworx account. Pack connections include shared gateway schemas in addition to the Inaturalist tools, which can increase the tool list and context size. The MIT-licensed package is an unofficial integration.