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

Query the data lake with SQL

Run one read-only DuckDB SQL statement (SELECT, WITH, FROM, DESCRIBE, SUMMARIZE) over the data lake's datasets, each a view by its name. Writes, other files and the network are refused. Local server only.

The tool is lake_query 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
sqlyesstring
limitnointeger

How a model reaches it

Ask in plain words. The model finds lake_query 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_query",
  "arguments": {
    "sql": "<string>"
  }
}

Words that find it

Besides its name, title and description, find_tool matches a question to lake_query on these words: sql, query, duckdb, data, lake, select, aggregate, join, window, analytics, local.

More

Nothing here is investment advice.