MotherDuck MCP

Name: MotherDuck MCP
Author: MotherDuck

by MotherDuck

Query and analyze data with MotherDuck serverless DuckDB and local files via SQL

database Python Intermediate Self-hostable Verified

⭐ 150 stars 📅 Updated: 1mo ago

View on GitHub View package

Description

MCP server for MotherDuck that enables LLMs to query and analyze data using serverless DuckDB in the cloud and local DuckDB databases. Seamlessly combines cloud-hosted data with local files — query CSV, Parquet, and JSON files on your machine alongside cloud tables in a single SQL statement. Powered by DuckDB's blazing-fast analytical engine, it supports listing databases, describing schemas, executing queries, and working with local files directly. Perfect for data analysts and engineers who want AI-assisted analytics without moving data between systems.

✅ Best for

Data analysts who want to query local files and cloud data together with AI assistance

⏭️ Skip if

You only need a traditional database connector without local file analysis

💡 Use cases

Running SQL analytics on local CSV, Parquet, and JSON files through natural language
Querying cloud-hosted MotherDuck databases from AI assistants
Combining local and cloud data in a single analytical query
Exploring and profiling datasets without writing SQL manually

👍 Pros

✓ Hybrid local + cloud queries — analyze files on disk alongside cloud tables
✓ Built on DuckDB's fast analytical engine with zero-copy data access
✓ Serverless architecture — no infrastructure to manage on the cloud side
✓ Simple pip or uvx install with token-based authentication

👎 Cons

✗ Requires a MotherDuck account and token for cloud features
✗ Local-only mode has limited functionality compared to cloud-connected usage
✗ Large result sets from analytical queries may consume significant LLM context

🔧 Exposed tools (4 tools)

Tool	Category	Description
list_databases	discovery	List all available databases in the MotherDuck workspace
describe_table	discovery	Get the schema and column details of a table
execute_query	query	Execute a SQL query against MotherDuck or local DuckDB
query_local_files	query	Query local CSV, Parquet, or JSON files using SQL

⚡ Installation

Prerequisites:

• python v3.10+
• MotherDuck authentication token
• API key required

Check Claude Code documentation to configure this MCP server.

💡 Tips & tricks

Set the MOTHERDUCK_TOKEN environment variable for authentication. You can query local files directly with SQL — try SELECT * FROM 'data.csv' LIMIT 10 to get started. Use the hybrid query capability to join local files with cloud tables seamlessly.

🔄 Alternatives

ClickHouse MCP

Query ClickHouse databases with natural language for real-time analytics and OLAP

Similar functionality

Tinybird MCP

Real-time analytics on Tinybird serverless ClickHouse — query, manage pipes, and publish APIs