the problem

GitHub Actions is slow. Jobs queue for 30-60 seconds. The dispatcher forgets to dispatch. Cold VMs spin up fresh every time. You wait a minute to find out you forgot a semicolon.

We built HorseCI because broken CI breaks flow. Jobs should run when you push, not when the dispatcher remembers.

immediate dispatch

We wrote a faster dispatcher. Webhook to worker in milliseconds. No queues. No forgotten jobs.

warm workspaces

Your worker stays on your machine. Dependencies are cached. The git repo is already there.

your code stays private

We dispatch webhooks. Your worker checks out code and runs builds. We never see your source.

what it is

HorseCI is a managed CI service. You run a worker binary on your machine. It connects to our dispatcher. When you push, we route the webhook to your worker immediately.

This is not self-hosted CI in the traditional sense. You do not manage the scheduler, the queue, or the GitHub integration. You run one binary. We handle the rest.

trade-offs

horseci

  • you provide the machine
  • your machine stays online
  • you update the worker
  • beta, rough edges

github actions

  • fully managed
  • always available
  • zero maintenance
  • proven, stable

when to use what

Use GitHub Actions if you want zero maintenance and don't mind waiting. Use HorseCI if you want fast feedback and don't mind running a worker. They can coexist.

try the beta