Run background workers on Vercel
Vercel Daemons continuously pull and execute work from queues, databases, and third-party systems. No HTTP server required.
For all your pull-based workloads
Pull tasks from queues
Daemons can poll messages from Vercel Queues, Celery, BullMQ, RabbitMQ, or Kafka consumers.Run workflows
Temporal or Cadence workers long-poll task queues for workflow and activity tasks.Host CI runners
Host GitHub Actions, GitLab, or CircleCI runners on Vercel to pick up jobs from your pipelines.
Deploy, monitor, and manage your daemons
Deploy daemons alongside your frontend and backend services. Set the working directory and start command for each daemon.
- 20:06:27ERROR celery: Task sync_inventory failed: TimeoutError
Monitor activity. Use built-in observability to spot issues and troubleshoot errors.
- orders-daemoniad1 · 3/5 ready5
Adjust capacity. Scale instances from the dashboard or API.
- SIGTERMinstance-1FinishedExitedinstance-2
Handle restarts gracefully. Clean up on termination signals before restarting.
Frequently asked questions
What are Vercel Daemons?
What are Vercel Daemons?
Vercel Daemons run background processes on Fluid compute. They are designed for workers that pull work from queues or other systems.
What can I run with Daemons?
What can I run with Daemons?
Any long-running, always-on process that pulls its own work, such as queue consumers, workflow workers, and CI runners.
How do Daemons scale?
How do Daemons scale?
You choose the number of running instances. You can adjust this manually or programmatically via API.
Do Daemons support persistent storage?
Do Daemons support persistent storage?
Persistent filesystem storage is not included in the initial release but will be added.
Can Daemons serve web requests?
Can Daemons serve web requests?
It is not possible to serve HTTP traffic from a worker in the initial release. We expect this capability to be added later.