@pipeworx/noaa-coastwatch

NOAA CoastWatch's satellite and model ocean record — sea surface temperature, chlorophyll-a, winds, currents and the full NDBC moored-buoy archive — read live from CoastWatch's two ERDDAP servers.

Part of Pipeworx — an MCP gateway connecting AI agents to 1683+ live data sources.

Tools

• noaa_coastwatch_search_datasets(query, protocol?, node?, limit?) — full-text search of the dataset catalogue. Start here: every other tool takes a dataset_id from these results.
• noaa_coastwatch_dataset_info(dataset_id, node?) — variables, units, axes, time coverage, licence. Call it before querying data; variable names are case-sensitive.
• noaa_coastwatch_tabledap(dataset_id, variables?, constraints?, node?, limit?) — rows from a tabledap dataset, e.g. hourly NDBC buoy observations back to 1970.
• noaa_coastwatch_griddap_point(dataset_id, variable, latitude, longitude, time?, depth?, node?) — one gridded value at one time/lat/lon, snapped to the nearest grid cell.

Auth

Keyless. Both nodes serve anonymously. A User-Agent is sent on every request.

Data sources

• https://coastwatch.pfeg.noaa.gov/erddap — node: "westcoast" (default). ERD/SWFSC: satellite SST, chlorophyll, winds, currents, and cwwcNDBCMet, the NDBC buoy archive.
• https://coastwatch.noaa.gov/erddap — node: "national". Operational VIIRS and ABI products.

The two hold different datasets. A dataset_id from one does not resolve on the other, and the error for that is "Currently unknown datasetID=…", which reads as a broken id rather than a wrong node.

Traps

Protocol-level traps live in shared/src/erddap.ts (shared with ioos-erddap) and are documented in full at the top of that file. The two that bite hardest here:

• ERDDAP reports "nothing found" as HTTP 404 with a plain-text body, not as an empty table — for a search that matches nothing AND for a tabledap query whose constraints exclude every row. The body is not JSON, so a naive parse loses the message entirely. That message is worth keeping: it names the dataset's real range ("No data matches time>=2050-01-01 because the numeric variable's source min=1970-02-26T20:00:00Z, max=2026-09-17T23:35:00Z").
• Variable names are case-sensitive and often UPPERCASE. NDBC water temperature is WTMP, air temperature ATMP, wind speed WSPD. A lowercase request is a 400 Unrecognized variable, which reads as "this buoy does not report water temperature".

One more, local to this pack: the two nodes do not agree on their search columns. The West Coast node returns 17 columns including "Accessible"; the national node returns 15 and omits it. Anything that indexed rows[6] would read the title on one and the ISO-19115 link on the other — a wrong answer, never an error. The shared client indexes by column name.

Quick Start

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

{
  "mcpServers": {
    "noaa-coastwatch": {
      "url": "https://gateway.pipeworx.io/noaa-coastwatch/mcp"
    }
  }
}

What this endpoint actually serves

tools/list at https://gateway.pipeworx.io/noaa-coastwatch/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 1683+ 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/noaa_coastwatch_search_datasets \
  -H 'Content-Type: application/json' \
  -d '{"query":"sea surface temperature","protocol":"griddap","limit":5}'

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

Or run it directly to confirm it starts:

npx -y @pipeworx/mcp-noaa-coastwatch

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 Noaa Coastwatch data" })

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

More

• Docs and guides
• pipeworx.io

License

MIT