Hosted Quick Start

Create your account

Go to app.useatlas.dev and sign up with your email or a social login provider (Google, GitHub). You'll be prompted to create a workspace — this is your team's home in Atlas.

Every workspace starts with a 14-day trial with full access to all features. See Billing & Plans for details on plan tiers after the trial.

Invite your team (optional)

Go to Admin > Users to invite team members by email. Invitees receive an email with a link to join your workspace.

Members can ask questions and share conversations. Admins can also manage connections, settings, and users. See Team Management for full role details.

You can always invite more people later — skip this for now if you prefer.

Prepare your database

Before connecting, create a read-only database user for Atlas. This is the only step that happens outside of Atlas.

CREATE USER atlas_reader WITH PASSWORD 'your-strong-password';
GRANT CONNECT ON DATABASE your_db TO atlas_reader;
GRANT USAGE ON SCHEMA public TO atlas_reader;
GRANT SELECT ON ALL TABLES IN SCHEMA public TO atlas_reader;
ALTER DEFAULT PRIVILEGES IN SCHEMA public
  GRANT SELECT ON TABLES TO atlas_reader;

CREATE USER 'atlas_reader'@'%' IDENTIFIED BY 'your-strong-password';
GRANT SELECT ON your_db.* TO 'atlas_reader'@'%';
FLUSH PRIVILEGES;

Atlas enforces SELECT-only queries at the application level, but a read-only user provides defense-in-depth.

Add the connection in Atlas

1. Go to Admin > Connections (/admin/connections)
2. Click Add Connection
3. Paste your connection string (e.g., postgresql://atlas_reader:password@host:5432/your_db)
4. Click Test to verify the connection — you'll see a confirmation with latency on success

Review the semantic layer

Atlas automatically profiles your database and generates a semantic layer — a set of entity definitions that describe your tables, columns, relationships, and business terms. The agent reads these before writing any SQL.

Go to Admin > Semantic Layer (/admin/semantic) to review what was generated:

• Entities — one per table, with columns, types, sample values, joins, and measures
• Glossary — auto-detected business terms and ambiguous field names
• Metrics — pre-built aggregation queries

You can edit entities directly in the web editor — add descriptions, adjust sample values, remove tables that shouldn't be queryable, or define custom measures. Better entity definitions produce better queries. See the Semantic Editor guide for details.