Quickstart

DuckHouse gives you a DuckDB database in the cloud that you query from DuckDB itself. There is no new client to learn and no driver to install: you ATTACH it and it behaves like another database in your session.

1. Create a database

1. Create an account and open the dashboard.
2. Choose Create database, give it a name, and keep the recommended region, US East, unless the programs that will query it run elsewhere. DuckLake databases always run there. The smallest size is fine to start with; you can resize later.
3. Copy the token you are shown. It is displayed once. If you lose it, create another from the database's page.

Provisioning takes about a minute. The database's page shows its connection address when it is running.

2. Install DuckDB

You need DuckDB 1.5.3 or newer. Your database runs DuckDB 1.5.5, and matching that version is the safest choice (see Versions and limitations).

# macOS and Linux
curl https://install.duckdb.org | sh

# or with Homebrew
brew install duckdb

duckdb --version

3. Attach and query

Start duckdb and attach your database using the address from its page and your token. DuckDB downloads its quack extension automatically the first time.

ATTACH 'quack:duckdb-a1b2c3d4e5.fly.dev:443' AS wh (TOKEN 'dh_your_token');

-- Create a table in the cloud and query it
CREATE TABLE wh.events AS
  SELECT range AS id, now() AS created_at FROM range(1000);

SELECT count(*) FROM wh.events;

That table now lives in your DuckHouse database. Close DuckDB, open it on another machine, attach again, and it is still there.

4. See what is in it

FROM wh.query('SHOW ALL TABLES');

Where to go next

Load your own data with one statement, connect from Python or Node.js, keep a Postgres database or Stripe account in sync, or hand your data to an AI agent over MCP.

Prefer a GUI to the terminal? DuckPlus is our native DuckDB IDE. Paste the same address and token and browse your tables, write SQL and see results (see Connecting).