Trigger.dev

Trigger.dev runs long-running background jobs written in your own codebase, without you managing the servers they run on. It is for the work that does not fit in a web request: sending an email campaign, processing video, running a compute-heavy job, or anything that takes longer than a serverless function will allow.

The central claim is that tasks have no timeouts. A job can run indefinitely, and a Checkpoint-Resume system means a long task can be suspended and picked up again rather than restarted from the beginning. Every task can be retried on failure, which is the part most homegrown queue setups get wrong. Concurrency is controlled with custom rules, so one noisy job cannot starve the rest, and tasks can be triggered in batches.

Because tasks are code in your repository, they are version-controlled like the rest of it. They can be run and tested locally before anything is deployed, then deployed to staging or production, across multiple regions including AWS US West in Oregon. Python scripts can be executed alongside the JavaScript. Human-in-the-loop workflows are supported for jobs that need somebody to approve a step partway through.