Connect a source.
Orchid talks to databases, warehouses, and dbt directly. Credentials stay in your OS keychain — they never leave your machine.
Add a connection
- Open the Integrations panel from the activity bar.
- Click + Add connection.
- Pick a connector type from the list.
- Fill in the host, database, user, and password fields.
- Click Test connection. If green, click Save.
The schema appears in the sidebar within a few seconds. You can now write SQL cells against it.
Supported connectors
Each connector has its own setup page with credentials and gotchas:
- Databases — PostgreSQL, MySQL, SQLite, Supabase
- Warehouses — Snowflake, BigQuery, Redshift
- Modeling — dbt (read models & tests)
Coming soon (v1.1 roadmap)
Read-only vs. write access
New connections start read-only. Granting write access is a separate, confirmed step — a lock icon appears on every cell using that connection so you always know which queries can mutate.
Agents draft queries. Read-only by default means a confused agent can't accidentally truncate your production table — it would hit a permission error before the data layer.
SSH tunnels & SSL
Both supported per connection. Open the connection's settings and toggle the relevant tab. SSH expects a private key (file or pasted); SSL accepts an optional CA cert.
What's next
With a connection in place, write your first SQL cell. See Notebooks. To version-control your project and push it to GitHub, see GitHub & source control.