frequently asked

A few questions, answered.

What is pondlog?

pondlog is an open-source nature data layer for developers. It wraps eight free public APIs (iNaturalist, eBird, USA-NPN, USGS, NOAA, Mushroom Observer, astronomy-engine, Trefle) and ships them as one CLI, six MCP servers, and a typed core library you can wire into your terminal, your agent, or your app. Garden planning (1,000-crop calendar, companion graph, USDA zones, climate types) lives in a separate product: CropGraph.

How do I install the CLI?

One line.

$ npm install -g pondlog
$ pondlog config set-location --lat 48.118 --lng -123.431
$ pondlog today

You need Node 18 or newer. pondlog today aggregates every source for your configured location into a single briefing. pondlog --help lists all 30+ subcommands.

How do I use the MCP servers?

Each MCP server runs over stdio and is installable with npx. Add an entry to your client’s MCP config and the tools appear in your agent. For Claude Desktop:

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

Swap @pondlog/mcp-pondlog for any of the six servers. The unified one (mcp-pondlog) exposes all sources through aggregate tools; the source-specific servers expose each source’s individual functions.

What data sources does pondlog aggregate?

Eight sources today:

  • iNaturalist, wildlife observations, all taxa, 100/min
  • eBird, bird sightings, hotspots, taxonomy, free key required
  • USA-NPN, phenology timing for bloom, leaf-out, flight
  • USGS Water Services, streamflow and gage height
  • NOAA CO-OPS, tide predictions, no auth
  • Mushroom Observer, mycology observations
  • astronomy-engine, moon, sun, planets, computed locally
  • Trefle, plant taxonomy with free token

Six of the eight require no auth at all. The two that do (eBird, Trefle) are free.

Is pondlog free?

Yes. All published pondlog packages and all six MCP servers are MIT licensed. CropGraph (the garden planning product) is also MIT. There is no paid tier and no plan to add one.

The upstream data providers each have their own terms. We pass through their auth requirements and rate limits unchanged.

How do I contribute a new source or MCP tool?

Source clients, MCP tools, and CLI commands all live in the pondlog monorepo. Clone an existing source-* package as a template for a new one, or edit an mcp-* package’s tools.ts to add a tool.

$ git clone https://github.com/andrewschristison/pondlog
$ cd pondlog
$ pnpm install
$ pnpm -r test

Crop calendar and companion graph contributions go to CropGraph instead. That is where the JSON fixtures and their JSON Schemas live.

What happened to the garden MCP?

Garden planning is now its own product. CropGraph has the garden planning MCP, CLI, and API: 1,000 crops, 333 companion relationships, 12 rotation families, USDA hardiness zones, and the six-type climate classifier. Same data, same authors, all offline, MIT.

If you previously ran npx @pondlog/mcp-garden, switch to npx @cropgraph/mcp. The ten tools kept their names and shapes.

What is the difference between pondlog and CropGraph?

pondlog is the open-source nature data toolkit: 8 source clients, 6 MCP servers, a CLI, and a typed core. You install it and run it.

CropGraph is the agricultural intelligence product: the 1,000-crop calendar, companion graph, USDA zones, and climate types, shipped as @cropgraph/core, @cropgraph/mcp, and a CLI. The hero of this page calls api.cropgraph.com live for your planting zone.

Use pondlog for wildlife, birds, tides, streamflow, phenology, mycology, and night sky. Use CropGraph for garden planning. They share an author and an ethos; they ship independently.

Where does the data come from?

The aggregated observations come from the eight public APIs named above. They are the canonical sources for their domains: eBird for birds, iNaturalist for everything else with eyes, Mushroom Observer for fungi, USA-NPN for phenology, USGS for water, NOAA for tides, astronomy-engine for the sky.

Garden data (crop calendar, companions, USDA zones, climate types) lives in CropGraph, curated by hand from USDA Cooperative Extension publications with citations preserved per entry.

Who built this?

Andrew Christison, on the Olympic Peninsula in Washington State. pondlog started as the data layer for a wall-mounted family kiosk and grew into a general-purpose nature data toolkit because every part of it was independently useful.

Find the source at github.com/andrewschristison/pondlog and the packages at npmjs.com/org/pondlog. Issues, pull requests, and crop submissions all welcome.