Features
Automation
Set a goal, execution option and trigger, inspect run history and pause work you no longer need.
Updated
Open Automation, or the related section inside a project. Try the request in a normal conversation before making it recurring.
Templates are editable starting points and never create or run tasks automatically. Configure daily ideas, account metrics sync, weekly reviews, or schedule checks before enabling them. Metrics sync requires the selected Agent to have authorized tools for the platform. Schedule checks do not approve or publish content.
For example, request a source-backed summary of a specific topic and ask for questions when evidence is missing. Broad instructions to operate an account every day give you little basis for checking completion.
Read run history, open the associated conversation and check outputs or pending confirmations. A successful trigger means execution was requested, not that every step succeeded.
Enabled means scheduled triggers are allowed. It does not mean a run is currently executing; check run history for execution progress.
Required approvals still apply. Automation does not imply every platform supports scheduled public publication.
Pause failing work, inspect the latest run and correct the source scope, execution option or instructions before a small retry. Archive automations you no longer use.
Keep Webhook addresses and credentials private. See Account and usage for charges.
Q: Why is there no final result after an automation triggers?
**A: ** Open the run record and its conversation. Check for pending approval, unavailable nodes, or execution errors. A successful trigger only starts the execution process.
Q: How should I adjust an automation that keeps failing?
**A: ** Pause recurring execution and inspect the latest failure. Correct the source scope, execution option, or instructions, then test a small run before resuming.