hatchet

Hatchet is a platform for running background tasks, AI agent workflows, and other jobs that need to happen reliably outside of a normal web request. When a user submits a form or triggers an action in an app, some work is too slow or too risky to do in real time. Hatchet provides the infrastructure to queue that work, run it in the background, retry it if something fails, and monitor what happened.

The platform supports applications written in Python, TypeScript, Go, and Ruby. You can use it as a hosted cloud service or run it yourself. It uses PostgreSQL, a common open-source database, as its underlying storage for both task state and observability data, which makes self-hosting more straightforward compared to systems that require separate specialized components.

Hatchet handles a range of task patterns. Simple one-off jobs can be queued and either forgotten or awaited. Jobs can be triggered by schedules (like cron), by incoming events, or by webhooks from external services. For more complex scenarios, tasks can be chained into directed workflows where each step runs after the previous one completes. There are also durable execution features for long-running workflows that need to survive server restarts or temporary failures.

The platform includes rate limiting (to avoid overwhelming external APIs), priority queues (so urgent jobs run before lower-priority ones), concurrency controls (to cap how many of a given job run simultaneously), and a web-based monitoring interface with alerting and logging. It also supports multi-tenancy, meaning one installation can serve multiple teams with separate data and user roles.

Hatchet is available as a managed cloud service with a free tier, or can be installed locally using its command-line tool and Docker.