Background Agents That Are Always On
Agents start from an event, a schedule, or a person, and work without anyone watching. When the team comes back, the results are waiting for review.
Started by an Event, a Schedule, or a Person
An automation pairs plain-language instructions with one or more triggers: a recurring schedule, a one-time run, or an event from GitHub, GitLab, Linear, Jira, PostHog, Vercel, or your own system.
New automation
When
- ScheduleWeekdays at 06:00
- EventPostHog alert firing
- MentionAsked in Slack
Do
Investigate and post a first analysis
Work Happens While the Team Is Away
Agents run on shared infrastructure, not on a developer's laptop, so work continues overnight and through weekends without anyone keeping a machine awake.
- 03:12Alert investigated, draft PR #1433 opened
- 04:40Flaky auth tests fixed
- 06:00Dependencies updated, tests green
Every Run Is Visible
Triggered runs appear as normal sessions, marked with where they came from, so the whole team can see what agents did and review the results.
Team sessions
- Automation: Nightly flaky test sweep3 tests fixed
- Webhook: Triage failed build #2291Fix opened
- Maria: Update checkout error copyIn review
Common Questions
What can start an agent?
A recurring schedule, a one-time run at a chosen time, an event from a connected system, or a person asking from the dashboard or a connected conversation.
Which systems can send events?
GitHub, GitLab, Linear, Jira, Asana, PostHog, and Vercel, plus any system that can send an authenticated request to Felan's webhook endpoint.
Do agents run on developers' machines?
No. They run on shared infrastructure, Felan's managed cloud or your own, so nothing depends on a laptop being awake.
Who reviews what agents did?
A person on the team. Every run is a session the whole team can see, and results come back for review.
Learn more: Use cases, Automations docs.
Keep Your Attention for the Hard Problems
Start with one repo and one recurring task.