From commit to production, without the wait
Relay is a CI/CD platform with pipelines as code, remote build caching, and preview environments on every PR. Queue-to-start in ~15s — ship with confidence, not queue anxiety.
Built for teams who ship every day
Caching, matrix builds, preview environments, and deploy gates — everything you need from push to production in one platform.
Remote Build Caching
Reuse layers across branches and runners. Builds finish 3.2x faster on average.
Matrix Builds
Test across OS, runtime, and dependency versions in a single pipeline definition.
Preview Environments
Spin up an isolated preview URL for every pull request — review before merge.
Cloud & Self-Hosted Runners
Run on managed Linux, macOS, Windows, and ARM — or bring your own hardware.
Secrets & OIDC
Scoped secrets, short-lived tokens, and OIDC federation — no long-lived keys in CI.
Pipeline Insights
Track build duration, queue time, and success rate across every team and repo.
pipeline: webon: [push, pr]jobs:build:cache: truerun: pnpm build
Pipelines as Code
Define stages, jobs, and gates in relay.yaml — versioned beside your application.
Deploy Gates & Rollbacks
Require approvals before production, then roll back to the last good deploy in one click.
The building blocks of a fast pipeline
From first commit to production deploy — the essentials, without the glue code.
Pipelines as code
Define stages, jobs, and gates in relay.yaml — versioned right beside your application.
Remote build cache
Reuse layers across branches and runners so builds finish in a fraction of the time.
Preview environments
Spin up an isolated, shareable URL for every pull request — automatically.
Matrix builds
Fan out across OS, runtime, and dependency versions from a single definition.
Secrets & OIDC
Scoped secrets and short-lived tokens — no long-lived keys living in CI.
Deploy gates & rollbacks
Require approvals before production, then roll back to the last good deploy in one click.
Relay AI handles the failures
you'd rather skip
Flaky test detection, auto-retry routing, and failure triage — built into every pipeline. Spend less time in logs, more time shipping.
Flaky test detection
Flag intermittent failures across recent runs
Failure triage
Summarize root cause from logs and traces
Routing auto-retry
CI result watcher
Retry failed jobs on merge to main?
Trusted at scale:
Builds per month
Median time to green
Uptime
Define, preview, and release
with confidence
Pipelines as codeKeep CI/CD beside your application in relay.yaml. Review pipeline changes in pull requests — same workflow, same audit trail.
Teams that stopped waiting on CI
Slow builds, flaky tests, deploy anxiety — sound familiar? Here is what engineering teams say after switching to Relay.
Start building freeOur main branch builds went from 18 minutes to under 6 with @relay caching. Queue time basically disappeared.
Flaky integration tests were killing us. @relay AI flagged the offenders and auto-retried the rest — we haven't babysat a red build in weeks.
Preview environments on every PR changed how we review. No more "works on my machine" before merge.
Our main branch builds went from 18 minutes to under 6 with @relay caching. Queue time basically disappeared.
Flaky integration tests were killing us. @relay AI flagged the offenders and auto-retried the rest — we haven't babysat a red build in weeks.
Preview environments on every PR changed how we review. No more "works on my machine" before merge.
Our main branch builds went from 18 minutes to under 6 with @relay caching. Queue time basically disappeared.
Flaky integration tests were killing us. @relay AI flagged the offenders and auto-retried the rest — we haven't babysat a red build in weeks.
Preview environments on every PR changed how we review. No more "works on my machine" before merge.
Our main branch builds went from 18 minutes to under 6 with @relay caching. Queue time basically disappeared.
Flaky integration tests were killing us. @relay AI flagged the offenders and auto-retried the rest — we haven't babysat a red build in weeks.
Preview environments on every PR changed how we review. No more "works on my machine" before merge.
Moved our relay.yaml into the repo and pipeline changes finally go through code review. Deploy anxiety is basically gone.
Matrix builds across Linux and macOS runners caught three OS-specific regressions last sprint. Worth it alone.
Deploy gates and one-click rollbacks mean on-call isn't a panic button anymore. @relay gave us a safety net we actually trust.
Moved our relay.yaml into the repo and pipeline changes finally go through code review. Deploy anxiety is basically gone.
Matrix builds across Linux and macOS runners caught three OS-specific regressions last sprint. Worth it alone.
Deploy gates and one-click rollbacks mean on-call isn't a panic button anymore. @relay gave us a safety net we actually trust.
Moved our relay.yaml into the repo and pipeline changes finally go through code review. Deploy anxiety is basically gone.
Matrix builds across Linux and macOS runners caught three OS-specific regressions last sprint. Worth it alone.
Deploy gates and one-click rollbacks mean on-call isn't a panic button anymore. @relay gave us a safety net we actually trust.
Moved our relay.yaml into the repo and pipeline changes finally go through code review. Deploy anxiety is basically gone.
Matrix builds across Linux and macOS runners caught three OS-specific regressions last sprint. Worth it alone.
Deploy gates and one-click rollbacks mean on-call isn't a panic button anymore. @relay gave us a safety net we actually trust.