Self-serve or managed — which one is for me?
If you want to drive it yourself, start on the self-serve tiers (Cloud Sandbox, free, or Self-Serve Pro at $49/mo): AI generates the tests, you review and run them in your CI, no engineer involved. If you want a forward-deployed engineer to own the suite — build it, keep it green, triage every failure — that's the managed tiers (Starter and up). Same platform underneath; the difference is who does the work.
Is the 30-day managed pilot really free? What's the catch?
Free, full-coverage, no credit card. The catch is honest: we run a limited founding cohort in 2026 because every managed pilot gets a real engineer's time. If the morning report doesn't sell itself, walk away and keep the tests — they're already in your repo.
What do you need from us to start?
Read access to your repo (via the scoped GitHub/GitLab app), a staging environment with a couple of synthetic test accounts, and a Slack channel invite. For managed plans you also send your critical flows in writing, a Loom, or by giving us staging access — your engineer maps them and proposes the suite back for approval.
Do I have to get on an onboarding call?
No — not below Enterprise. Onboarding is async: send your critical flows in writing or on a recording and your engineer takes it from there. Enterprise is the one tier that includes live onboarding calls if you want them.
How fast until I see results?
Five working days from intake to your first 9am report in Slack — pass/fail across disciplines, bugs already filed, a clear "safe to ship?" verdict. The full breakdown is on the How It Works page.
How does pricing work?
Flat monthly per maintained tests, not per run. Runs are always unlimited in your CI — you never pay per execution. Every number is public; you never need a demo to find out the price. Annual billing saves 15%.
What are the tiers?
Three self-serve with no engineer (Cloud Sandbox, free; Self-Serve Pro, $49/mo; Team, $149/mo) and four managed with a forward-deployed engineer: Starter ($499/mo, pooled engineer, up to 60 tests), Growth ($1,799/mo, dedicated engineer, up to 200 tests), Scale ($2,499/mo, up to 500 tests with Performance + Security included), and Enterprise (custom). Full detail and INR pricing is on the pricing page.
Will my price change mid-contract?
No. Flat monthly fee for tests under management. The only thing that moves your invoice is you — adding tests beyond your cap ($1.75 / ₹150 per test/month, no upgrade required) or adding an add-on. Everything costs what the page says.
What if I need more tests than my plan's cap?
Add maintained tests anytime at $10/test/month (₹700) — each one is maintained by your engineer. Once you'd be adding a lot, the next tier is better value (and Growth gives you a dedicated engineer), so the pricing nudges you to upgrade rather than pile on add-ons. Your cap is the number of maintained tests, never a limit on how often they run.
Are Performance and Security extra?
On Starter and Growth they're optional add-ons with published prices — Performance ($179/mo) and Security ($249/mo). On Scale both are included at no extra cost. On Enterprise everything is scoped into the engagement. No hidden metering.
Can I cancel anytime?
Yes — cancel anytime, effective at the end of the current billing month. No termination fees, no clawbacks. Annual plans refund unused months pro-rata, minus the 15% annual discount you received.
What's the difference between Cloud Sandbox and Self-Serve Pro?
Cloud Sandbox is free forever: AI generates up to 15 UI tests/month so you can get the feel in our cloud sandbox, plus unlimited CI runs once you wire your pipeline. Self-Serve Pro ($49/mo) unlocks UI + API generation (200/month), self-healing locators, flake detection, and 3 projects.
Why is the cloud capped but CI unlimited?
Running browsers on our infrastructure costs us money; running them in your CI doesn't. The cloud sandbox exists so you can try it in minutes — the product is built to live in your pipeline, where runs are free and unlimited.
Do self-serve plans include an engineer?
No — self-serve is fully self-driven, no engineer and no call. When you want a human to own the suite and triage failures for you, that starts at the managed Starter tier.
What does "pooled" vs "dedicated" mean?
On Starter, your suite is owned by a shared pool of engineers — you get real human ownership without paying for a full dedicated hire. From Growth up, you get a genuinely dedicated engineer whose name is on your account and your morning report.
What does the engineer actually do?
They review and approve every AI-generated test before it enters your suite, rewrite the ones that assert the wrong thing, triage every overnight failure (real bug vs flake vs environment), write the "safe to ship?" verdict, and file verified bugs with repro steps. AI handles volume; the human handles judgment.
Who at QAShift can see my code?
A named engineer and one backup, both identified in your onboarding doc. Access is via scoped app permissions with SSO and hardware-key MFA, and it's logged. Details are on the Security page.
If AI writes the tests, what are the humans for?
Verification. AI generates and maintains the suite; on managed plans a named engineer reviews every generated test before it merges and re-runs every failure before it reaches your report. You never triage flaky-test noise — that's the core of the product.
What if a test breaks because the UI changed, not because of a bug?
Self-healing locators absorb most UI drift automatically — you can redesign a flow and not touch a test file. When intent genuinely changes, your engineer updates the test the same day, included in the flat fee, not billed as a change request.
How do you tell a flake from a real failure?
Failures are classified automatically — real bug, flaky environment, or regression — each with a confidence score. High-confidence bugs are filed straight to your tracker; ambiguous ones are re-run and investigated by your engineer before they ever reach your report. The goal is zero false alarms.
What are the predictions, and are they gated by tier?
You get flake, regression, and confidence signals, plus two Beta predictions (cross-dimension correlation and a deploy confidence score). They're gated on data, not tier — once you have ~20 deploys of history they light up. Below that threshold you see an honest "not enough data yet" state, at every tier.
Do you block merges?
Yes — merge-blocking PR checks are available from Starter up. A red gate means a human-verified bug, never a false alarm, with repro steps attached to the PR.
How do bugs reach my team?
Auto-filed into Jira or Linear with repro steps, video, trace, and the suspect commit — before the 9am report lands. Standup starts with everything already triaged.
What can you actually test?
Five disciplines: UI functional testing, API/integration testing, accessibility, performance (load testing at production traffic levels with regression alerts), and security (OWASP Top 10 scanning, dependency vulnerability audits, and static analysis every build). UI and API are core; performance and security are add-ons that become included at Scale.
Do you test mobile apps?
Yes — native iOS/Android mobile app testing is included from Growth up, if your product ships an app. It runs on simulator/emulator by default, with real-device testing available at Scale.
How does accessibility coverage work?
Continuous accessibility scanning runs on every deploy and is never counted against your test cap. It's included from Starter. For deploy-over-deploy accessibility regression monitoring in the PR itself, see Clearpath.
Do I own the tests?
Yes. Every test is committed to your repo from day one — Playwright for web, API, visual and AI-feature evals (native mobile authored by your engineer) — and runs in your CI with no QAShift dependency. The specs and the Page Object Model structure are yours to keep.
Is any of it open source?
Yes — the core runner (@qashift/verity-core) is MIT-licensed. You're never locked into a proprietary format you can't run yourself.
What happens to the tests if I leave?
Nothing — they're already yours and already in your repo. Leaving costs you the verification service and the engineer, not the asset. Your suite keeps running in your CI.
Does my code leave my repository?
No. We work inside your repo via scoped app permissions — read on application code, write only on the test directory and CI config. No mirrors, no forks. Full details on the Security page.
Do you sign NDAs and DPAs?
Yes — a mutual NDA before we see anything (our template or yours), and a DPA for GDPR / India DPDP included with every plan.
What compliance do you support?
GDPR and India DPDP alignment on every plan, WCAG 2.1 AA as the accessibility baseline, and OWASP Top 10 scanning. Enterprise adds SOC 2 / GDPR support, SSO/RBAC, and a self-hosted runner.
What are Verity and Clearpath?
They're the products QAShift Labs ships from the same engine behind our managed QA. Verity is the AI-augmented, multi-discipline testing platform (UI + API, self-healing, runs in your CI). Clearpath is per-deploy accessibility regression monitoring that tells your team exactly what broke since the last deploy, right inside the pull request.
Should I use a product or the managed service?
Use Verity or Clearpath if you want to self-serve and own the workflow. Use the managed tiers if you'd rather a forward-deployed engineer run it all for you. Many teams start self-serve on Verity and move to managed as coverage grows.