// ci/cd test automation
Block bad deploys with one CI step.
One API call from your pipeline fires a full regression suite on cloud workers — generated Playwright tests, accessibility, SEO, security and performance audits — with results back in your PR thread.
// the flow
PR merge → tested → comment in under 6 minutes.
You merge a PR
The merge fires your CI workflow as usual.
CI calls AegisRunner
A single trigger-token API call. We pick up the new commit and start a run.
We run the suite
Generated Playwright tests + audits (a11y, SEO, security, perf) run in parallel on cloud workers.
PR gets the verdict
Pass/fail comment on the PR with deep links into the failure trace. CI gate blocks merge on red.
// integrations
Pipeline-agnostic — works with what you already use.
GitHub Actions
Trigger token + PR comment with results.
GitLab CI
Same trigger token, drop into .gitlab-ci.yml.
Vercel
Deploy hook — every preview deploy gets a scan.
Netlify
Build-success webhook fires the suite.
Custom webhooks
HMAC-SHA256-signed payloads for your own pipeline.
Anywhere via curl
`curl -X POST /api/v1/trigger` — works from anything.
// one call from your ci
It really is one step.
- name: Regression test via AegisRunner
run: |
curl -X POST https://app.aegisrunner.com/api/v1/projects/$PROJECT_ID/trigger \
-H "Authorization: Bearer $AEGIS_TOKEN" \
-H "Content-Type: application/json" \
-d '{"branch":"${{ github.ref_name }}","wait":true}' \
--failthat's it. exit code = build verdict.
// faq
Frequently asked
How fast does a CI run finish?
Typical SaaS suite (50 tests, 3 browsers) finishes in 3–6 minutes on parallel cloud workers. Replay-cache shaves another 30–60% off on re-runs against unchanged pages.
Does AegisRunner block deploys on a failure?
You choose. The trigger-token endpoint returns the run result; your CI can fail the job on red. We also post a PR comment with the verdict so reviewers see it without leaving GitHub.
What about parallel runs / concurrency?
Included on every paid plan. Pro runs 8 parallel workers per project; Business runs 32. No per-minute fees, no per-test pricing.
Can the same suite run locally?
Yes — Playwright export on every plan. The generated suite is plain TypeScript, runs in your own `npx playwright test` against your own browsers. Useful for debugging individual failures locally.
Wire it up in five minutes
Sign up, mint a CI token, paste one step into your workflow. We'll handle the rest.