Automations

Actions

Each automation can have one or more actions that run when the trigger fires. This article covers the configuration options for each action type.

Run tests

The Run tests action executes individual tests or test groups. Select the tests and groups you want to run from the multi-select picker. You need at least one test or group.

Branch selection

Choose which branch the tests should run against. You have two options:

  • From trigger - use a branch from the event that triggered the automation. For example, if the trigger is "Branch created", you can select "Created branch" to run tests on the newly created branch. If the trigger is "Branch merged", you can choose either "Source branch (from)" or "Target branch (to)". If the trigger is "Branch removed", you can use the removed branch value.

  • Project branches - select a specific branch from your project, such as Live or any development branch.

For deploy-triggered automations, select your Live branch. For branch-based triggers, use the "From trigger" options to automatically test the affected branch. Manual / API and Recurring automations do not provide a trigger branch, so choose one of your project's branches instead.

Deploy agent

The Deploy agent action kicks off an AI agent with a prompt you write. Configure the model, prompt text, and where the final result should be sent.

Template parameters

For branch-based triggers, you can insert dynamic values into your prompt using double-brace syntax. Click any available parameter below the prompt field to insert it, or type it manually.

Available parameters depend on the trigger type:

  • Branch merged - {{branchFrom.id}}, {{branchFrom.name}}, {{branchTo.id}}, {{branchTo.name}}

  • Branch created - {{branch.id}}, {{branch.name}}

  • Branch removed - {{branch.id}}, {{branch.name}}

  • Manual / API, Live deployed, and Recurring - no trigger parameters are available.

Parameters are highlighted in your prompt as you type so you can see which parts will be replaced at runtime.

For example, a prompt like "Review the changes on {{branchFrom.name}} before it merges into {{branchTo.name}}" would resolve to "Review the changes on feature/payments before it merges into main" when a merge triggers the automation.

The agent runs using the connected providers of the user who created the automation. This is shown in the configuration panel.

Send result

Choose where the agent's final response should be forwarded after the run completes.

  • Email - send the result to one or more email addresses.

  • Webhook - POST the result to an HTTPS endpoint, with optional custom headers.

  • Slack - forward the result to a Slack channel after you connect Slack in Buildprint.

Send webhook

The Send webhook action sends a POST request with a JSON payload to a URL you specify. The URL must use HTTPS.

Headers

Add custom headers if the receiving endpoint requires authentication or specific content types. Buildprint always includes Content-Type: application/json automatically. Click Add to add key-value header pairs.

Payload

The payload is generated automatically and includes:

  • event - the trigger type (for example, "manual", "live_deployed", or "branch_merged").

  • automation - the automation ID and name.

  • project - the project ID and app name.

  • params - trigger-specific parameters. For branch events, this includes the branch ID and display name. For Manual / API, live deploy, and recurring triggers, this is empty.

  • timestamp - when the event fired (ISO 8601).

A live preview of the payload is shown below the URL field so you can see exactly what will be sent.

Testing webhooks

Click Send test request next to the payload preview to fire a test POST to your URL with the preview payload. This lets you verify your endpoint is receiving and parsing the data correctly before the automation goes live. The button is only available when you've entered a valid HTTPS URL.

The test request sends the preview payload with placeholder values, not real event data. Use it to verify connectivity and parsing, not to test actual automation logic.

Was this helpful?