Release QA for teams who actually ship

Green CI isn’t the same as a safe release.

ReleaseQA reads your test runs, scores every build, and flags the flaky and regressing tests hiding under a passing pipeline — so you know which green is actually green.

Free forever tier · 14-day Pro trial · No credit card

releaseqa — main — last 6 builds
Quality score
89 +3
~
Flaky tests
5 quarantined
!
Regression
checkout.spec
Ready to ship. No blocking regressions. Flaky tests quarantined.
rqa verdict: GO
3
npm reporters
JUnit XML
any other runner
1 line
in your CI config
$0
to start, no card

What you get

Four signals. One verdict. Zero noise.

Built for teams that have outgrown 'CI is green, ship it.'

Quality score

A single number for ship-readiness.

Every build gets a weighted score across pass rate, flake rate, coverage drift, and regression signal. Know at a glance if main is safe.

Flake detection

Catch unreliable tests before they rot CI.

ReleaseQA fingerprints failures over time, flags intermittent tests, and lets you mark them for quarantine — so CI stops lying.

Regression radar

See which specs broke and which PR broke them.

Commit-linked regressions, ownership attribution from CODEOWNERS, and blame that points at the change — not the test.

Works where you work

Plug into any CI, any framework, in minutes.

Playwright, Cypress, Jest, Vitest, and a generic CLI. GitHub Actions, GitLab, CircleCI, Jenkins, Buildkite. Zero vendor lock-in.

What your team actually sees

The comment CI wishes it could leave.

ReleaseQA reads every run and drops a single, actionable verdict on the PR — before anyone hits merge.

#1428feat(checkout): add apple-pay supportCI / testpassedreleaseqa / verdictHOLD
releaseqa[bot]botcommented on this pull request
HOLD — don’t merge yet
CI is green, but 2 signals say this release isn’t safe.
score: 68
  • !
    Regression: checkout.spec.ts › applies discount code
    Passed on main, failing since c7a1f9e · owner @checkout-team
  • ~
    Flaky: auth.spec.ts › remembers sessions (passed 3/5 on retry — hiding real failures)
    Suggested: quarantine · see 7-day flake history
releaseqa · run #4,281

One verdict. Commit-linked. Owner-tagged. No dashboards to check.

How it works

Three steps. About the time it takes to make coffee.

No agents to install. No sample data to seed. Point it at a real build and see a real score.

  1. 01

    Install the reporter

    Reporters for Playwright, Cypress and Jest. Any other runner sends JUnit XML through the CLI.

    npm i -D @releaseqa/reporter-playwright
  2. 02

    Pipe it into your CI

    Add one line to your existing workflow. No agents, no sidecars, no new infrastructure.

    npx @releaseqa/cli report --format junit --file results.xml
  3. 03

    Ship on a verdict, not a hunch

    Every build gets a score and a GO / HOLD verdict. Wire it into PR checks, Slack, or deploy gates.

    rqa verdict: GO (score 89, +3)
Get your first verdict

Integrations

Fits into the stack you already have.

ReleaseQA observes — it doesn't replace your CI, your runner, or your alerting. Wire it in once.

Test frameworks

Available today

PlaywrightCypressJestGeneric CLI

On the roadmap

VitestMochaPytest
CI / CD

Available today

GitHub ActionsGitLab CICircleCIJenkins

On the roadmap

BuildkiteAzure PipelinesBitbucket Pipelines
Signals & alerts

Available today

SlackGitHub checksWebhooksEmailREST API

On the roadmap

LinearPagerDutyJira

Runner not listed? Drop us a line — we add integrations most weeks.

Straight answer

Early access, built by a QA engineer.

Not because it runs your tests better — because it tells you the truth about them. Here is the truth about us.

Who built it

One QA engineer who spent years writing Playwright suites for other companies and got tired of green pipelines that lied.

Where it is

Early access. The reporters are on npm, the dashboard is live, and the founder reads every signup. No case studies yet — we will publish the first one with the team's permission.

What we will not claim

No invented customer counts, no borrowed logos, no compliance badges we have not earned. If a number is on this site, it came from the product.

Want to talk to the person who built it before you install anything? Email the founder.
example — last 4 releasestrend: improving

Pricing

Free while you evaluate. Fair when you scale.

Every tier includes the quality score, flake detection, and the PR verdict. Pick the plan that matches your run volume.

Free

Free

Get started at no cost

  • 100 test runs / month
  • 1 project
  • Quality score & GO / HOLD verdict
  • Flaky test detection
Start free
Most popular

Pro

$29/mo

14-day free trial — no credit card

  • 1,000 test runs / month
  • 5 projects & 3 seats
  • Everything in Free
  • 30-day history
See full plan

Team

$79/mo

For serious engineering orgs

  • 10,000 test runs / month
  • 25 projects & 10 seats
  • AI failure analysis
  • Quality gates & PR checks
See full plan
Enterprise · SSO, audit log, custom retention
High-volume ingestion, dedicated support, annual contracts.
Talk to us →

Full feature comparison · see /pricing →

Why it exists

We got tired of shipping on vibes.

ReleaseQA was built by engineers who spent one too many Friday evenings rolling back a release a green CI swore was fine.

A passing pipeline is not a promise.

Most CI systems tell you 'all green' while a flaky test hides the one that actually broke. ReleaseQA separates signal from ceremony.

Quality is a number, or it isn't real.

If your team can't answer 'is main shippable right now?' in one glance, you're deploying on memory. We turn that into a score.

Your release cadence shouldn't depend on luck.

Confident teams ship more often. We built ReleaseQA so you can be one of them — without buying an entire new testing platform.

Frequently asked

The questions every QA lead asks.

Didn’t see yours? support@releaseqa.com.

  • Where is my test data stored?
    Run metadata (pass/fail, duration, test names, commit SHAs) is stored on AWS in US-East. We never ingest your source code. EU data residency is available on Enterprise. See privacy and DPA.
  • Can I self-host ReleaseQA?
    Not today. ReleaseQA is cloud-only. If on-prem is a hard requirement for your procurement, reach out — we're gauging demand for an Enterprise self-host tier.
  • What exactly does the reporter send?
    The reporter ships test names, durations, pass/fail status, failure messages, stack traces, and git metadata (SHA, branch, PR number). Screenshots and video traces are opt-in per project. No request/response bodies are captured unless you configure an explicit attachment.
  • Does it work with my framework?
    Playwright, Cypress, Jest, and Vitest have first-class reporters. Everything else (Mocha, Pytest, Go test, RSpec, JUnit XML) runs via npx @releaseqa/cli. If you emit JUnit XML, you’re supported today.
  • What happens when I hit the free-tier run limit?
    Ingestion keeps working — you just stop seeing new runs in the dashboard until the next calendar month or until you upgrade. We don't delete historical runs. Free retention is 30 days; Pro is 90; Team is 1 year.
  • How is this different from my CI system?
    CI tells you if tests passed. ReleaseQA tells you whether the passes are trustworthy. A test that fails 3/5 times but gets retried to green is hiding a real regression — ReleaseQA flags that before you merge; CI celebrates it.
  • Do you support monorepos?
    Yes. One org can hold many projects; each reporter run tags its project, so quality scores, flake history, and verdicts are scoped per package or app.
  • Is there a free tier for OSS projects?
    Yes — public GitHub repos get unlimited runs on the Free plan. Ping oss@releaseqa.com if you need a higher project cap.