canli-mcp / The desk, built into canli-mcp

Save a tool's table into the data lake

Run a tool and save its table (columns and rows) into a dataset in the data lake on this machine, as a new vintage stamped with the time saved. Every free source the server reads can feed the lake this way. Needs DuckDB: npx -y canli-mcp lake setup. Local server only.

The tool is lake_save in canli-mcp 0.22.0, free and MIT-licensed, with no account and no API key. It is part of canli-mcp itself, so it is there as soon as canli-mcp is added.

Arguments

As the server declares them
ArgumentRequiredType
datasetyesstring
nameyesstring
argumentsnoobject
pathnostring

How a model reaches it

Ask in plain words. The model finds lake_save with canli-mcp's find_tool, reads its arguments with describe_tool if it needs to, and runs it with run_tool. The call has this shape, with your values in place of the angle brackets:

{
  "name": "lake_save",
  "arguments": {
    "dataset": "<string>",
    "name": "<string>"
  }
}

Words that find it

Besides its name, title and description, find_tool matches a question to lake_save on these words: data, lake, save, store, parquet, duckdb, local, warehouse, cache, history, dataset, vintage, point, in, time.

More

Nothing here is investment advice.