Agentic QA PlatformNew

E2E tests that write, run, and
fix themselves.

Cairn connects to your GitHub repo, maps your app's flows, and ships Playwright coverage that repairs itself when your code changes.

Map flows
Cairn Agent
U1
U2
Generate tests
Checkout flow
User authentication
Payment webhook
File issues
Payment webhook timeout
#42 · filed 2m ago
Login redirect fixed
#41 · resolved
The problem

E2E tests are a full-time job. Nobody signed up for that.

Writing tests takes as long as building the feature

Every new page, form, or flow means hours of Playwright boilerplate — selectors, waits, auth setup, assertions. Most teams skip it.

Tests break every time the UI changes

A CSS refactor, a renamed button, a new onboarding step — any change can break 20 tests. Fixing them is grunt work that nobody wants to own.

Failures are mysteries

A red CI check with a cryptic timeout error. No video, no trace, no context. You re-run it locally, it passes. You merge anyway.

Auth flows are a nightmare to automate

OAuth redirects, magic links, session cookies, signed tokens — getting a headless browser past the login page is half the battle.

No one watches the test suite

Tests pass when you write them and break a month later. Nobody notices until a customer reports the bug the test was supposed to catch.

Test code drifts from app code

Pages are added, routes renamed, components refactored — the test suite doesn't know. It tests yesterday's app against today's code.

Cairn does all of this for you. Automatically.

How it works

Four steps. Zero test code in your repo.

Cairn lives outside your codebase. Connect, map, generate, run — everything stays in Cairn's dashboard.

01

Connect your repo

Install the GitHub App. Cairn gets read access to your code and write access to issues. Nothing else.

02

Map every flow

Cairn reads every page in your app, systematically discovers user flows, and groups them by feature area.

03

Generate tests

One click per flow or bulk-generate. Cairn writes deterministic Playwright specs with auth handling baked in.

04

Run & fix automatically

Schedule runs against staging or prod. Failures auto-file GitHub issues with video, traces, and a diagnosis.

Features

Everything you need. Nothing you don't.

AI mapping

Page-by-page mapping

Cairn reads the source of every page in your app — not just filenames. It discovers view flows, form submissions, CRUD operations, and error states systematically.

Test gen

Playwright test generation

Each flow becomes a self-contained .spec.ts file. No shared fixtures, no imports from your repo. Cookie injection, selector strategies, and wait patterns built in.

Issues

Auto-filed GitHub issues

When a test fails, Cairn drafts a natural-language issue with the action, likely cause, affected files, and a link back to the run. Deduplicates flaky failures automatically.

Automation

Scheduled + manual runs

Set a cadence per environment — hourly, daily, weekly. Or hit Run anytime. Inngest orchestrates each flow as an independent step with automatic retries.

Performance

Parallel execution

Configure max concurrency per environment. A 150-test suite at concurrency 5 finishes in ~15 minutes instead of 75. Each flow gets its own browser context.

Debugging

Video + Playwright traces

Every run captures a video recording and a full Playwright trace — DOM snapshots, network, console, action timeline. Debug failures without reproducing them.

Drift detection

Change-aware sync

When code merges to main, Cairn compares the diff against your flows. Changed pages are flagged, new pages are surfaced, stale flows are marked — before they break.

Auth

Auth bypass

Cairn detects your auth library, explains how to mint a long-lived session token, and injects it per-environment. Authenticated tests skip the login UI entirely.

Iterate

Chat agent

Ask Cairn to add a flow, edit one, rewrite a test, or explain a failure. The chat reads your repo files, proposes changes with Accept/Discard cards, and persists everything.

Changelog

What's shipped. What's next.

Next
Up nextComing soonPLANNED
  • PR-triggered runs — automatic suite on every preview deployment
  • Neon DB branching — isolated data per PR preview environment
  • Auto-fix PRs — Claude drafts a fix when a test breaks
0.5
v0.5Apr 2026LATEST
  • Auto-filed GitHub issues — LLM-drafted body with action, cause, affected files, and dashboard deep link
  • Issue deduplication — flaky failures bump a counter instead of spamming duplicates
  • Issues tab with accordion detail view, date filter, and one-click open on GitHub
  • Change-aware sync — push to main flags affected flows automatically
0.4
v0.4Apr 2026
  • Parallel execution with per-env configurable concurrency
  • Playwright trace capture alongside video recording
  • Auth detection + session token management per environment
  • Chat teaches auth setup — library-specific guidance with non-expiring tokens
0.3
v0.3Apr 2026
  • Runs tab with scheduled + manual suite execution via Inngest
  • Multi-environment support (staging, prod, preview)
  • Apps/worker with Playwright Docker image for production deploy
  • Schedule presets — hourly, every 4h, daily, weekly
0.2
v0.2Apr 2026
  • Page-by-page systematic flow mapping (pre-reads all source)
  • Per-flow test generation with parallel support
  • Chat agent with readFile, listDirectory, and proposal cards
  • Flow + test + chat persistence across sessions
0.1
v0.1Apr 2026
  • Dashboard with Map / Tests tabs
  • GitHub App integration for repo access
  • Playwright test runner with video capture
  • Better-Auth GitHub OAuth login

Stop writing tests. Start shipping.

Cairn maps your app, writes the tests, runs them on schedule, and files issues when things break. You ship.