
PikoCI — The CI/CD that grows with you
How hard can it be to build a CI/CD system? Concourse CI has the right model but the operational overhead is brutal. That question stuck with me long enough that I started building it. What kept me going was realising I needed it for my own side projects too: games and open source tools that require custom environments GitHub Actions can't provide.
The design goal was something that grows with you. A binary and a pipeline file is all you need to start: runs entirely in memory. Add SQLite for persistence. Add Postgres and distributed workers when you scale. Never have to migrate or reconfigure.
Key things:
- Single binary, zero setup, in-memory by default
- Run pipelines locally: pikoci run --pipeline-config pipeline.hcl --job test runs any job on your laptop. No server, no push, no waiting.
- Services: start any process before your tasks, stop it after, guaranteed. No Docker-in-Docker.
- Four sourceable abstractions: resource types, runners, services, secret backends. All defined in HCL, all pullable from a URL. Write it once, host it anywhere, reference it by URL.
- HCL pipelines: Terraform-style syntax
- Public pipelines: share status without an account
PikoCI deploys itself. Live at ci.pikoci.com/teams/main/pipelines/pikoci, no login needed.