Analyze the codebase
Discover journeys such as account creation, checkout and account management from your repository.
CodeRabbit Verify
Test runs are free during public betaCodeRabbit Verify is an agentic QA platform that generates user journeys and tests them against your app. Debug failures with recordings, screenshots, and logs, then send findings to your coding agent to fix.
AI coding agents produce more changes than teams can test by hand. Someone still has to run the app, walk through user journeys, and report bugs back to the agent. CodeRabbit Verify takes that first testing pass so your team can inspect failures and incomplete checks before shipping.
CodeRabbit Verify uses your PR diff and codebase context to generate test scenarios and run them against your application.
Apply SAVE20 at checkout
PR test scenarios
PR changes · code · documentation
Canvas tote
1 × $100.00
Recorded charge does not match the discounted total.
CodeRabbit Verify analyzes your repository and proposes user journeys, including the user’s goal, steps and expected outcomes. Review and refine the generated journeys, then run them against your application on demand or on a schedule.
Discover journeys such as account creation, checkout and account management from your repository.
Refine each journey’s actions and expected outcomes before approving and saving it.
Test the running application on demand or on a schedule, then inspect the results of each run.
acme/shop
app/
app/checkout/
app/checkout/page.tsxMatched
app/api/orders.tsMatched
app/signup/page.tsx
app/account/
app/account/settings.tsx
README.md
Routes and application behavior
Generated user journeys
User goal
Complete an order with a valid discount.
Steps to test
Expected outcome
The order total and simulated charge both equal $80.
Each finding includes the expected behavior, the observed result and supporting evidence. Inspect the failure before sending the findings to your coding agent for a fix.
Follow test scenarios in a live browser, then inspect recordings, screenshots and logs.
Compare the expected user journey with results from testing the application with recorded evidence of the test.
Send the bugs and blockers to CodeRabbit Agent for a fix with relevant context automatically included.
Recorded browser steps
00:24Discount is missing from the recorded charge
Checkout displays an $80 total, but the application records a $100 simulated charge.
Developer guidance
Keep the displayed total and recorded charge consistent. Check the amount after a page reload.
Repair context
Finding, tested revision, environment and evidence
Markdown
Reproduce: apply SAVE20 to a $100 order and complete checkout.
Expected: order total = $80; simulated charge = $80.
Observed: order total = $80; simulated charge = $100.
Context: tested revision 9f2c4a1, sandbox and execution evidence.CodeRabbit Agent prepares a sandbox, installs dependencies, configures supporting services and starts your application. Reusable setup instructions and environment snapshots reduce preparation for later runs. Testing happens in CodeRabbit’s sandbox, separate from production.
Install dependencies
Project packages
Configure services
Local database and supporting services
Start the application
Check application readiness
Save the prepared environment
Setup instructions and snapshot
Generate scenarios for a pull request, test the running application and inspect the results.
Run Verify from a pull request @coderabbitai verify
If the repository is not configured, follow the setup link in CodeRabbit’s reply to prepare its test environment.