# DuckHouse

> DuckHouse is managed DuckDB in the cloud. Each database is a DuckDB 1.5.5 server on its own machine that you attach from any DuckDB client over Quack, DuckDB's own client/server protocol, with a REST API and an MCP server for automation and AI agents.

- Connect from DuckDB 1.5.3 or newer: `ATTACH 'quack:duckdb-a1b2c3d4e5.fly.dev:443' AS wh (TOKEN 'dh_your_token');`. The address and a token are on the database's page in the dashboard.
- Some statements do not work through an attached database yet. `FROM wh.query('…')` runs SQL on the server, where everything works; see "Querying and loading data".
- REST API: https://duckhouse.co/api/v1. MCP server: https://duckhouse.co/api/mcp (OAuth or an organization API key).
- Every page below is Markdown. All of the documentation in one file: https://duckhouse.co/llms-full.txt
- Agent skills: `npx skills add https://duckhouse.co` installs the skill listed below.

## Start here

- [Quickstart](https://duckhouse.co/docs.md): Create a DuckDB database in the cloud and query it from the DuckDB CLI on your own machine, in about two minutes.
- [Connecting](https://duckhouse.co/docs/connect.md): Attach a DuckHouse database from the DuckDB CLI, Python, Node.js, the DuckPlus IDE or anything else that embeds DuckDB.
- [Querying and loading data](https://duckhouse.co/docs/querying.md): What works through an attached database, how to list tables, and how to move data in and out.

## Your databases

- [Sizes, idling and lifetime](https://duckhouse.co/docs/databases.md): Machine sizes and regions, stopping when idle, temporary databases, and how your data is kept safe.
- [Tokens and API keys](https://duckhouse.co/docs/tokens.md): Database tokens for DuckDB clients, and organization API keys for the REST API and MCP.
- [Connections](https://duckhouse.co/docs/connections.md): Keep a copy of your databases and the tools you run on in your warehouse, on a schedule.
- [Source setup guides](https://duckhouse.co/docs/sources.md): Shopify, Square, Google Analytics, Meta Ads, Google Sheets, Notion and GitHub: setup, caveats and tables.

## Automation

- [REST API](https://duckhouse.co/docs/api.md): Create, query and manage DuckHouse databases over HTTP with an API key: run SQL, read schemas and set up Connections.
- [MCP for AI agents](https://duckhouse.co/docs/mcp.md): Connect Claude, ChatGPT, Codex, Cursor and other MCP clients with OAuth or an API key.

## Reference

- [Versions and limitations](https://duckhouse.co/docs/limits.md): The DuckDB version we run, what Quack cannot do yet, and what to expect from DuckDB 2.0.

## Product

- [Pricing](https://duckhouse.co/pricing): Per-minute compute while a database runs, storage at cost, and a cost estimator.
- [Connectors](https://duckhouse.co/connectors): Every source DuckHouse syncs into a database, and how often.
- [DuckPlus](https://duckhouse.co/duckplus): DuckPlus, the free, open-source Mac IDE for DuckDB from DuckHouse. Connects to DuckHouse databases over Quack. Latest .dmg: https://github.com/wes/duckplus/releases/latest/download/DuckPlus.dmg

## Agent skills

- [duckhouse](https://duckhouse.co/.well-known/skills/duckhouse/SKILL.md): Work with DuckHouse (duckhouse.co), managed DuckDB databases in the cloud. Use when connecting DuckDB, Python or Node.js to a DuckHouse database over Quack (ATTACH 'quack:…' with a dh_ token), writing SQL that has to work through an attached remote database, loading or exporting data, or automating databases with the DuckHouse REST API or MCP server (dhk_ API keys).

## Optional

- [Changelog](https://duckhouse.co/changelog): What has shipped, newest first.
- [All documentation](https://duckhouse.co/llms-full.txt): Every page above in one Markdown file.