Repros records the evidence while you test: console errors, failed requests and screenshots, each timestamped against the exact change and build under test. No more “can you reproduce that?”
No card required. Testing needs no SDK: the extension watches the browser tab, not your codebase.
Step 1 of 3. Press Apply. The promo code is broken on this build.
Your cart
REC Session time 00:14
0 errors0 failed0 notes
01 — The problem
A tester finds something broken and describes it from memory. A developer then spends the next twenty minutes trying to get back to the exact state where it happened, often without success.
Testerthe promo code thing is broken on checkout, discount doesn't show up
Developercan you repro? what browser, what code, logged in or guest?
Testerlet me try again, i think it was chrome. give me a sec
Developerany errors in the console when it happened?
Testeri didn't have devtools open, sorry
Everything a developer needs, before they have to ask.
02 — Alternatives
Bug reporters, session replay and test management each cover a piece. None of them gives you a deliberate test session for each change that captures its own technical evidence as it runs.
Jam, Marker.io · One-off reports
Good for capturing a single bug with some technical context, after the fact.
What's missing: No structured session, so there's no record of everything that happened while testing change X on build Y.
LogRocket, PostHog · Continuous recording
Records everything, all the time, in production. Built for debugging real users after something has gone wrong.
What's missing: No way to say “I'm testing X now”, no pass or reject sign-off, and no assigned scope.
TestRail, Zephyr · Test cases and checklists
Organises what to test and tracks pass or fail, but testers still attach evidence by hand, if at all.
What's missing: No automatic evidence. Console logs, network requests and screenshots are all manual work.
A structured, assignable test session for each change, with technical evidence captured automatically inside the session. Not bolted on, and never reconstructed later.
03 — How it works
Step 1 · Dashboard
Point it at a domain pattern, such as app.example.com, or *.preview.example.com for pull request previews. Then add your builds and the changes to test (features, fixes or tweaks), each with a short description.
Step 2 · Extension
They sign in once. Work assigned to them appears in the extension popup, with a desktop notification. There's nothing to configure and no separate app to keep open.
Step 3 · Extension
The tester opens an assigned change, already pinned to the right build. One click opens the build, the toolbar appears and capture starts straight away.
Step 4 · Extension
Console errors, uncaught exceptions, network requests, clicks and page changes are logged in the background. Testers press N to add a note with a severity (a screenshot is attached automatically) and tick off the checklist as they go.
Step 5 · Extension
Pass or reject, with notes. A rejected session goes back for a retest as soon as a new build is ready.
Step 6 · Dashboard
Deduplicated errors, failing endpoints and automatically written steps to reproduce sit at the top, with the full timeline underneath. Copy it as a Markdown bug report or share a read-only link, so the developer gets the exact session rather than a paraphrase.
04 — Features
Console errors with the line that logged them, uncaught exceptions, and API calls, including 4xx and 5xx failures with their redacted payloads. Plus the browser, operating system and viewport. All captured the moment it happens, with no copying and pasting.
Every note can carry an automatic screenshot for context and a severity level, so the dashboard triages itself.
Developers define pass/fail items and testers work through them live in the toolbar, so a sign-off means more than “seemed fine”.
Assign a specific change and build to a specific tester. It appears straight away in their extension popup, with no hunting through a list to find what to test.
Set permissions for each member of your organisation: test, manage content, manage users and view analytics. Restrict a contractor to only the work assigned to them.
Every click and page change becomes a numbered step, with errors marked where they happened. Export everything as a Markdown bug report, copy any request as cURL, or share a read-only link.
/checkoutSPRING25 into “Promo code”/api/cart/apply-promo, TypeError in the consoleClaude Code, Cursor or any MCP client can read a rejected session's full evidence, see the screenshots, and send it back for retest — no copying logs into a chat window.
05 — Roadmap
The evidence trail shouldn't stop at sign-off. We're building two integrations next, so a rejected session goes straight into the tools your team already uses.
Coming soon
When a session is rejected, the right channel hears about it straight away, with the severity, the notes and a link to the full evidence. Not just “something broke”.
Coming soon
Turn a rejected session into a GitHub issue in one click, with the summary, severity and notes filled in and a link back to the full evidence. The issue closes itself when the retest passes.
Is there another tool you'd like us to prioritise? Suggest an integration.
06 — Pricing
£29/month
£79/month
Custom
Every plan starts with a free 30-day trial, including 5 seats. No card required.
07 — FAQ
Repros is a Chrome extension and dashboard for structured manual QA testing. Each test session automatically captures console errors, failed network requests and screenshots as they happen, tied to the exact change and build being tested. Your bug reports arrive with their own evidence instead of a best guess at what went wrong.
Any manual testing of a web app in Chrome: exploratory testing, regression checks before a release, verifying a bug fix, or user acceptance testing with people outside the engineering team. Each session is tied to one change and one build, so the evidence always comes with its context.
Bug reporting tools capture one bug at a time, and test management tools track pass or fail but leave the evidence to the tester. Repros combines a structured, assigned test session with automatic capture of console errors, network failures and screenshots, and ends every session with a pass or reject sign-off.
Not for QA testing. There's no SDK to add to your codebase: the extension watches the browser tab being tested, not your application's code. The optional Repros SDK is only for Customer Sessions, where you send a customer a one-time link to capture a repro on your own app, with no extension to install and no Repros account needed.
Starter is £29 a month for up to 5 testers. Team is £79 a month for up to 20 testers, with unlimited changes and session retention. Enterprise is priced individually, with negotiated seat counts and single sign-on (SSO). Every plan starts with a free 30-day trial, and no card is required.
Yes. The extension is free for testers to install and use. Pricing is per workspace, not per install.
Repros is currently a Chrome extension. Support for other browsers isn't on the roadmap yet.
Yes. Repros runs a hosted MCP server, so any MCP-compatible client can read a rejected session's full evidence — report, timeline, screenshots — and, with a write-scoped token, comment on it or send it back for retest. Create a personal API token from Account settings to connect it.
Slack notifications and a GitHub Issues integration are both in development. A rejected session will be able to notify your team or become a GitHub issue that links back to its evidence, and that issue will close itself when the retest passes.