Functions

Some work cannot run in the browser. A Stripe secret key or an email API key would be visible to anyone who opens the developer tools, and a browser tab cannot run a job at 3 a.m. or react when another user changes a row. Functions are where that code goes. It runs on the server, reads secrets your app never sees, and starts on its own when its trigger fires.

1. 1

   Your agent writes the function

   From a description, or from your own source code. A function whose code does not compile and load is never saved.

2. 2

   You choose what starts it

   A user signing up; an insert, update or delete on a table; a schedule (every 15 minutes, or a 5-field cron expression); an HTTP request to /api/v1/{projectId}/fn/{name}; or only when you run it by hand.

3. 3

   It runs on the server

   It can read encrypted environment variables you set for the project, so secret keys never reach the browser.

4. 4

   Every run is logged

   The console shows each run with its result, duration, log lines and error. You can test-run a function from the console before relying on it.

Once your coding agent is connected, ask it:

When a user signs up, send them a welcome email through Resend.

Every night at 2:00, delete carts older than 30 days that never became orders.

• Runs are counted per month, and the allowance depends on your plan. See Pricing.
• Turning a function off stops its trigger without deleting it. Deleting one from your agent waits for a person to approve it.
• There are no SQL functions (rpc()). Logic that would live in a stored procedure runs here instead.