Warehouse MCP server

Read from the package source without running it. The installed server may list more.

• column_stats

  Return min, max, null count, distinct count, and (for numeric columns) average for one column.

• count_rows

  Return the row count for a table. Cheap (single COUNT(*)), and answers 'how big is this?' before the agent decides whether to scan it. Especially useful on cloud warehouses where bytes scanned costs money.

• describe_table

  Return column names, types, and nullability for a table or view.

• find_columns

  Search column names across the warehouse using a SQL LIKE pattern (use % for wildcards). Returns matching (schema, table, column, type) tuples. Optionally scope to one schema. On BigQuery, schema is required because INFORMATION_SCHEMA is per-dataset.

• get_foreign_keys

  Return declared foreign-key relationships (from_schema/from_table/from_column → to_schema/to_table/to_column). Optionally scope to a schema or a single table. BigQuery only returns informational FKs that were explicitly declared.

• get_view_definition

  Return the SQL body of a view. Useful when business logic lives in views — the model can reason about a view's semantics by reading its definition.

• list_schemas

  List all schemas (or BigQuery datasets) visible to the configured warehouse role.

• list_tables

  List tables and views in a given schema.

• querywrite action

  Execute a single read-only SELECT against the configured warehouse. Returns columns and rows. A LIMIT (or FETCH FIRST on Oracle) is auto-applied if the SQL omits one.

• sample_table

  Return up to N sample rows from a table (capped at 100).

• search_value

  Find a literal value across the text columns of a table (exact match). Scans up to 10 text columns and returns up to 100 hits.

• time_series

  Group a table by a date/timestamp column into buckets (hour/day/week/month/quarter/year) and return a count or aggregate per bucket. Saves the agent from constructing dialect-specific date truncation. Returns rows ordered chronologically.

• top_values

  Return the top-K most frequent non-null values in a column with their counts.

• Code scan46 source files scanned25/25
• –Live reliabilityno gateway calls yet and no remote to proben/a
• –Tool poisoningtools not inspected (local package is not executed); not countedn/a
• Auth qualitystatic API keys via environment variables6/15
• Maintenancelast push 101 days ago8/15
• Maintainer identityregistry namespace matches repository owner; GitHub account older than a year8/10

Safety score

72/100

higher than 30% of 802 Databases servers

Tools

13

the typical Databases server has 8

Tools that change things

8%

typical: 10%. Fewer write tools means less an agent can break

Downloads a week

52

more than 40% of 324 Databases servers with a package

Warehouse: common questions

Is Warehouse MCP server safe?

Mostly: it is graded B (72/100). Read the Warehouse safety report

How do I install Warehouse?

It runs on your machine. Copy the Claude Code, Cursor, VS Code or Claude Desktop config from the install section.

Does Warehouse need an API key?

Yes. The registry entry asks for MCP_API_KEYS, MOTHERDUCK_TOKEN, PG_PASSWORD, ORACLE_PASSWORD.

Is Warehouse maintained?

The last commit was 137 days ago (2026-06-18). The latest release is v0.4.2.

What can I use instead of Warehouse?

Servers from other publishers that do the same job: Faucet MCP server, DBHub MCP server and Appwrite MCP server.

How does Warehouse compare with other Databases MCP servers?

safety score 72/100 (higher than 30% of 802 Databases servers); tools 13 (the typical Databases server has 8); tools that change things 8% (typical: 10%. Fewer write tools means less an agent can break). Full comparison.