list_data_sources

Discover all data sources attached to a project, including table names and column schemas. Use this BEFORE query_data to learn what tables and columns are available — it avoids the need for a dummy exploration query. Returns both uploaded Parquet files and linked Google Sheet snapshots. For each source, the ‘tables’ field lists the exact DuckDB table names to use in query_data SQL (e.g. SELECT * FROM <table_name>). The ‘schema’ field maps each table name to an ordered list of columns with their names and dtypes, giving enough metadata to write correct SQL. Google Sheet sources include a ‘snapshot_taken_at’ timestamp; call connect_google_sheet to refresh stale snapshots before querying. Intended flow: list_data_sources → inspect tables/columns → query_data. Returns: project_name (str); sources (list) — one entry per data source with fields: source_name (str), source_type (‘parquet’ | ‘google_sheets’), tables (list[str] — DuckDB table names), snapshot_taken_at (str ISO 8601 UTC | null), schema (dict[table_name, list[{name, dtype}]]), data_source_id (str UUID | null — for linked Google Sheets), spreadsheet_id (str | null — Google Sheets document ID; use this with write_gsheet or pass source_name instead); total_count (int) — number of data sources.

read-only · idempotent

Parameters

Name	Type	Required	Default	Description
project_name	str	Yes	—

Returns

Returns a structured object. Top-level fields:

• project_name (string)
• sources (array<object>)
• total_count (integer)

See the response example below for the full payload shape.

Example

Tool Call

{
  "name": "list_data_sources",
  "arguments": {
    "project_name": "forecasts"
  }
}

Response

{
  "project_name": "forecasts",
  "sources": [
    {
      "source_name": "actuals",
      "source_type": "parquet",
      "tables": [
        "actuals"
      ],
      "snapshot_taken_at": null,
      "schema": {
        "actuals": [
          {
            "name": "month",
            "dtype": "timestamp[us]"
          },
          {
            "name": "revenue",
            "dtype": "float64"
          }
        ]
      },
      "data_source_id": null,
      "spreadsheet_id": null
    },
    {
      "source_name": "headcount_sheet",
      "source_type": "google_sheets",
      "tables": [
        "headcount_sheet__roles",
        "headcount_sheet__hires"
      ],
      "snapshot_taken_at": "2026-04-28T10:00:00+00:00",
      "schema": {
        "headcount_sheet__roles": [
          {
            "name": "role",
            "dtype": "string"
          }
        ]
      },
      "data_source_id": "88888888-8888-8888-8888-888888888888",
      "spreadsheet_id": "1aBcDeFgHiJkLmNoPqRsTuVwXyZ0123456789ABCDEFG"
    }
  ],
  "total_count": 2
}

Notes

• Module: Data
• Requires authentication via API token or Auth0 JWT.