CodeRabbit Verify
Intro
CodeRabbit Verify is the agentic QA product CodeRabbit put into public beta on 18 November 2026. The official post Introducing CodeRabbit Verify describes it as a way to test your running application, check behavior changed by a pull request or across the codebase, and catch functional bugs or broken workflows before they reach customers.
The stated problem is volume: coding agents produce more changes than a team can click through by hand. Someone still has to deploy a test environment, walk the affected flows, and send bugs back to the agent. Verify is meant to take that first pass and return evidence a person or a coding agent can act on. The product page is CodeRabbit Verify; the docs are at CodeRabbit Verify docs.
This entry records the new product name CodeRabbit Verify. It should not be merged with CodeRabbit’s existing code-review comments, or with CodeRabbit Agent’s repair workflow.
Advertisement
Explain
Verify has two run types. A pull request run writes scenarios from the PR’s commits, related source, existing tests, and documentation, and can also use context from a CodeRabbit review of the latest commit. Expected results come from intended behavior, docs, and existing tests, so a new implementation does not get to define its own correctness. Results are posted on the PR and can be inspected in Change Stack. A user-journey run analyzes the repository and proposes reusable journeys in plain English—who the user is, what they want, the steps, and the outcomes that must hold. After you revise and approve them, you can run selected journeys on demand or on a schedule. The docs are explicit: pull request runs do not use saved journeys, and journey runs do not comment on pull requests.
Testing happens in CodeRabbit’s isolated sandbox, not in production. CodeRabbit Agent can inspect the project, prepare the sandbox, start the app, and confirm it is ready, then save reusable setup and startup instructions plus an environment snapshot for later runs. Failures keep recordings, screenshots, and logs. Application bugs are reported separately from blockers such as missing credentials that stopped a test from finishing. Findings can be sent to CodeRabbit Agent for a fix, or exported to another coding agent.
Per the docs, Verify is available on CodeRabbit Cloud for GitHub Cloud, GitLab.com, and Bitbucket Cloud. It tests web apps and services, not native mobile apps. On pricing during public beta, the blog says PR test runs, full-repo journey runs, and fixes with CodeRabbit Agent are all free; the product-page FAQ says PR and journey runs are free, while Agent fixes are billed as Agent minutes. Check the live official pages before you rely on either statement.
Usage
Read Introducing CodeRabbit Verify and the docs at CodeRabbit Verify, then add the repository in the CodeRabbit app and prepare its test environment. On a pull request you want tested, comment @coderabbitai verify. If the repository is not configured yet, follow the setup link in CodeRabbit’s reply. You can also turn on automatic runs when a pull request opens or gets new commits.
For regression coverage, generate user journeys from the codebase, review the steps and expected outcomes, approve them, and run them on demand or on a schedule. When a test fails, inspect the recording and screenshots before sending the finding to CodeRabbit Agent or exporting it. The docs also mention automatic issue creation in a connected Jira or Linear tracker for scheduled runs.
Verify is not a replacement for an existing test suite or a QA team; CodeRabbit describes it as a first pass that works alongside them. If you only want CodeRabbit’s usual review comments on a pull request, you do not need this new Verify product name.