All guides

10 min read

Connect a database

Give an agent Postgres, MySQL, or ClickHouse. It introspects the schema, writes real SQL, and reasons over the results in its own sandbox.

  1. 01

    Create a read-only role first

    Before you touch Helios, create a database role that can read the schemas a task needs and nothing else. A reporting agent with a read-only credential has no path to a write, because the permission is enforced by your database — which is exactly where that decision belongs.

  2. 02

    Add the connection

    From Settings → Integrations, add a Postgres, MySQL, or ClickHouse connection with a host, port, database, and credentials. Everything is stored envelope-encrypted and decrypted only when a run needs it.

  3. 03

    Reach a private database through an SSH tunnel

    When the database sits on a private network, configure an SSH tunnel on the connection. Agents connect through the bastion you specify, so nothing has to be exposed to the public internet to make this work.

  4. 04

    Ask a question in plain English

    Run a manual agent that asks something you already know the answer to. It introspects tables, columns, types, and keys, writes the SQL, runs it, and explains the result. Comparing its answer against the one you knew is the fastest way to see whether it understood your schema.

  5. 05

    Read the SQL it wrote

    Every run records the queries the agent ran and the rows they returned, so a report always traces back to the exact SQL. When a number looks wrong, the query that produced it is one click away.

  6. 06

    Schedule the task

    Attach a cron trigger and the report shows up before standup, every morning. Memory carries what the agent learned about your schema across runs, so the fifth report starts smarter than the first.

Try it in your workspace.

Start free, connect a tool, and describe the first task in plain English.