For QA engineers

A performance regression you can actually point at.

Load testing already covers your server. Redline covers the other half: the device. Run a repeatable stress test on real hardware, keep the baseline, and turn “it feels slower” into a number you can ticket.

Run a baseline test

A baseline you can reproduce

Run the same five-subsystem stress on any device and record the Redline Index. Every future run compares against that baseline, so regressions show up as a number — not a vibe.

Catches what load testing misses

Load tools tell you the server held up. Redline tells you the cheap phone didn't. Together they cover both ends of the delivery path.

Shareable reports, no setup

Send a stakeholder a Redline report instead of a screenshot of a flame chart. The verdict, cohort and fixes are written in plain engineering language.

CI budgets, shipping today

POST a URL and a performance budget to the Redline audit endpoint. It returns 412 when the budget is breached, so a regression fails the build instead of reaching production.

Gate a release on a budget

Add this to any pipeline. A non-zero exit means the delivery budget was breached.

curl -sS -X POST https://redline.lovable.app/api/public/v1/audit \
  -H 'content-type: application/json' \
  -d '{"url":"https://example.com","budget":{"minScore":80,"maxTtfbMs":600}}' \
  -o report.json -w '%{http_code}' | grep -q 200