How Tracefox tests your checkout, automatically.
Four layers working together: a Chrome extension (Pro and Business) or your store URL gives you a test, a scheduler queues it at the cadence you set, a Playwright executor replays it in headless Chromium on the viewport you chose for that test, and a notifier alerts you when a step fails. No code, no Playwright scripts, no QA team needed.
The four-layer test pipeline
Every test run flows through the same pipeline. Each layer has one job and hands clean output to the next. When a scheduled run fails and a confirming re-run fails too, you are alerted, not on Monday morning when revenue's already gone.
Record
Click through your checkout in our Chrome extension (Pro and Business). It captures clicks, typing, dropdown choices, checkboxes, scrolling and page changes on the main page, and stops at payment. Or generate the test from a Shopify, WooCommerce or BigCommerce store URL, or build it by hand.
Schedule
Pick a cadence per test: manual, weekly, daily or hourly, depending on plan. Different tests can run at different cadences. Runs on our own BullMQ + Redis job infrastructure.
Replay
Real headless Chromium runs the steps on the viewport you choose. A screenshot is taken after each step and, on Pro and Business, a video of each run is kept, pass or fail (except runs with sensitive inputs).
Alert
When a failure is confirmed by a re-run, you get an email (plus Slack on Pro and Business) with the broken step, the error, and the screenshot. At most one alert per test per hour, so stuck tests don't spam.
And we keep watching your scripts, too.
On top of checkout tests, every plan also runs a separate scan-and-diff pipeline against your site to catch silent script, cookie, and consent-banner regressions, the kind of stuff that breaks attribution without breaking checkout. The rest of this page is how that pipeline works under the hood.
A real browser, every time
Screenshot tools and string-match monitors miss the changes that actually matter. Tracefox uses headless Chromium to load each page the way a real visitor's browser would, running its scripts, showing its consent banner, and following redirects, so the snapshot reflects what a first-time visitor's browser loads.
- Scripts & pixels: third-party tags captured with src and load order (presence, not whether events send the right data)
- Cookies: full set after the page stabilizes, including those set after consent
- Consent banners: detected CMP, visible vs hidden state, decline / accept outcomes
- Policy pages: privacy and cookie policies tracked for material content changes
- Screenshots: full-page captures kept for each scan as visual evidence
acme-store.comScan completed 2m agofbq('init', '123456...'), last seen 3 days ago
__gcl_au, set by Google Ads conversion tracking
+2 paragraphs under 'Third-party services'
Not every change is a fire
A missing Facebook Pixel is a client-emergency. A new font file is noise. Tracefox classifies every detected change into HIGH, MEDIUM, or LOW severity, so you get paged for things that matter and batched digests for the rest.
- HIGH, tracking pixels disappearing, consent banners breaking, privacy policy removed
- MEDIUM, new marketing scripts, new cookies, CMP configuration changes
- LOW, minor content edits, asset updates, non-behavioral styling changes
Replay your clients' checkouts on autopilot
Beyond watching what changes, Tracefox can re-run full checkout flows on a schedule. Shopify, WooCommerce, BigCommerce, or any custom funnel. Each run keeps a screenshot of every step (and a video on Pro and Business), so you see which step failed and on which viewport.
- Record once in your browser with the Chrome extension (Pro and Business), replay on a schedule on our infrastructure
- Desktop, mobile, or tablet viewport per test (mobile and tablet are emulated in Chromium, on Pro and Business)
- Video + step screenshots for every run, so you can see where a failed run broke
- Coupon codes, custom data and success conditions, test the exact flow you care about
Product → cart → shipping → payment page.
A screenshot after each step, plus a video of each run on Pro and Business (not for runs with sensitive inputs).
Pick the cadence per test, plan limits apply.
Know on the next scheduled run if the Add to Cart button stopped working.
Route the signal where your team already lives
HIGH-severity detections alert as soon as a scan finds them. Everything else rolls up into digests so inboxes stay clean. Client reports are on every paid plan, with your agency's branding on Pro and Business.
Per-change and weekly digest options, with additional recipients per workspace.
Slack
Block Kit-formatted messages with direct links to review each detection. Pro and Business.
Webhooks
Signed payloads (HMAC-SHA256) for your own integrations. Business plan.
PDF reports
Timestamped client reports on paid plans, with your agency logo and name on Pro and Business.
Designed for people who manage many sites
Every part of the platform is built on the assumption that you're not looking at one website, you're looking at fifty.
Workspaces + teams
Invite teammates into a shared workspace. Everything is scoped per-workspace, no cross-client leakage.
White-label
Your agency logo and name on PDF reports and public status pages, on Pro and Business. Clients see your brand, not ours.
History retention
7 days on Free up to 90 days on Business, cleanup is automatic.
Audit log
Every mutation recorded with IP + actor, so you can answer 'who changed this and when?' instantly.
Webhooks + Chrome extension
Pro unlocks the extension recorder. Business adds signed webhooks, so failures reach your own tools.
Plan-gated resources
Downgrading automatically pauses excess sites and freezes excess tests, nothing silently exceeds your plan.
Every day without monitoring is a risk.
The Free plan includes 1 checkout test and 1 site monitored monthly. No credit card, no migration, no setup call.