to_pandas takes variables and constraints alongside the dataset ID, so what arrives is the slice you asked for rather than a file you then filter down. The other tools are the order of operations around it: find a server, search it, read the dataset metadata — which is where the variable and constraint names come from in the first place.
A client for ERDDAP data servers: dataset search, metadata lookup, a starting list of servers, and a data pull returned as CSV in a pandas DataFrame.
- Dataset search against whichever ERDDAP server URL you point it at
- Detailed metadata for one dataset by ID, with the protocol selectable
- A list of well-known ERDDAP servers for when you do not have one in mind
- A download narrowed to chosen variables and constraints, handed back as a pandas DataFrame
The URL of an ERDDAP server you want to query — every tool but the server list takes one per call — and a Python environment able to receive a pandas DataFrame.
One command — npx mcp-remote https://erddap2mcp.fly.dev/
