@pondlog/mcp-mushroomobserver

@pondlog/mcp-mushroomobserver

The first dedicated Mushroom Observer MCP server. Gives any MCP-aware AI client (Claude Desktop, Cursor, Continue, custom agents) five tools for fungal observations, mycology taxonomy, and region discovery from the largest dedicated mycology platform (500,000+ observations with vote-weighted ID confidence scores).

No API key required. Mushroom Observer is a public scientific community platform.

Part of pondlog, a toolkit for stitching together public nature APIs (iNaturalist, eBird, USGS, NPN, NOAA, SunCalc, Mushroom Observer).

What is Mushroom Observer?

Mushroom Observer is a community of mycologists, mushroom hunters, and scientists who document fungal observations with photos and vote-weighted identification confidence. Unlike a generic biodiversity platform, MO is mycology-first: its taxonomy, voting system, and geography are designed around fungi.

Install / run

The server is published to npm and runs via npx (no install step):

npx -y @pondlog/mcp-mushroomobserver

It speaks MCP over stdio. The server is internally rate-limited to 20 req/min (1 every 3 s) to respect Mushroom Observer's posted limits.

Configure

Claude Desktop

Edit ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows):

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

Restart Claude Desktop. The five tools will appear in the slash menu.

Cursor

Edit ~/.cursor/mcp.json:

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

MCP Inspector (debug)

npx @modelcontextprotocol/inspector npx -y @pondlog/mcp-mushroomobserver

Tools

All tools are read-only. Coordinates use WGS84 decimal degrees. Region strings use Mushroom Observer's "City, County, State, Country" suffix model.

Tool	What it does
get_recent_fungi	The headline tool. "What's fruiting near here in the last N days?" Accepts coords+radius OR region suffix.
search_observations	Rich filters: bbox, region, taxon name, date range, has-images, confidence threshold.
get_observation	Single observation, high detail: namings, votes, full image set, comments.
search_fungal_names	Substring search over MO's 100,000+ fungal name index, with rank filter.
search_regions	Discover MO's location-name suffixes (e.g. "Olympic National Park, Clallam Co., Washington, USA").

Key gotchas

• Mushroom Observer's /locations endpoint has no name filter. Use search_regions (which scans observations and harvests unique location names) to discover well-formed suffixes.
• Geography is suffix-based. Region strings match "ends with X"; pass the most-specific tail you want. "Washington, USA" matches all of Washington; "Clallam Co., Washington, USA" matches just Clallam County.
• Coverage is uneven. Pacific Northwest, Northern California, the Northeast US, and Western Europe are dense. South America, Africa, and most of Asia are sparse.
• Confidence is in [-3..3]. Vote-weighted across observers. 1.0+ is a reasonable filter for "agreed-upon ID"; 2.0+ is strong consensus; negative numbers mean the consensus name is disputed.
• Some observations have hidden GPS (gps_hidden=true). The tool still returns the location-name centroid in that case so the LLM has a reference point.

Example prompts

After configuring, try:

• What fungi have been observed near Port Angeles, Washington in the last 60 days?
• Find all chanterelle observations in Marin County, California with photos.
• Show me observation #187521 in detail.
• What's the difference between Cantharellus formosus and Cantharellus cibarius? Find recent examples of each.
• Where in Olympic National Park have mycologists observed Mycena haematopus?

Notes for AI agents

• For "what's fruiting near me?" call get_recent_fungi first. It pages through MO results internally and accepts either coords+radius or a region suffix.
• For "look up a name" use search_fungal_names. It returns the scientific name, author, classification, and a link to the MO name page.
• For broader region discovery use search_regions to enumerate location-name suffixes. This is the only path to discovering MO's geography programmatically.
• All tools return the same Result<T> envelope. Success returns { ok: true, data }; error returns isError: true with { ok: false, error: { source, message, statusCode? } }.

License

MIT. See the root LICENSE.