
Revix AI: Code Review Based on Your Team’s Rules
Repository-aware AI code review for your team’s standards

What does Revix actually do?
Revix AI is an AI code review service for professional software teams. Connect a supported code host and it checks new pull requests against conventions it finds in your repository. In everyday terms, it gives each proposed change an automated first review, with comments that explain the issue and the rule behind it. It is particularly relevant when AI coding assistants have increased the amount of code your team must inspect.
Turn your repository conventions into review rules
Revix looks beyond the changed lines to build repository context and derive review rules. You can edit, add or disable those rules, and they evolve after merges. This makes the workflow useful for checking whether a change fits the structure and conventions already used by a team.
Comments distinguish issues from smaller suggestions with labels such as issue and nitpick. Each explains the problem, a possible fix and the relevant rule. A short PR summary describes the change and marks it Has issues or Ready to merge. Treat that label as an AI assessment that still needs validation.
From a new pull request to actionable feedback
- Sign in with GitHub, Bitbucket or GitLab using an account authorized to connect the intended repositories.
- Select the repositories to review. For GitHub, install the app on your account or an organization you manage and restrict repository access as needed.
- Inspect the extracted rules and adjust conventions that do not reflect the team’s intentions.
- Open a new pull request, or a merge request on GitLab. Revix starts from provider webhooks and posts a summary and inline findings.
- Check each finding against the code and tests, push your fixes and inspect the thread state. Revix can resolve its own thread when a later commit fixes the issue. Keep the normal human merge decision.
Cloud platforms, permissions and operating limits
The documented integrations support GitHub.com, Bitbucket Cloud and GitLab.com. GitHub Enterprise Server, Bitbucket Data Center/Server and GitLab Self-Managed/Dedicated are not currently supported. Reviews run outside your CI pipelines, so no Actions, Pipelines or GitLab CI/CD minutes are needed.
The GitHub App reads code and metadata and can write checks and PR content. GitLab group connections require Maintainer or Owner access; comments use the connecting account. The vendor describes broad programming-language support, but that is not a guarantee of equal review quality for every stack.
Free public repositories and paid team plans
Checked on 29 October 2026: the Open Source tier covers public repositories without a subscription. Private-repository use moves to a paid plan after the 14-day trial, which does not require a credit card.
The official pricing page shows Starter at €20 per seat per month and Pro at €40 when billed yearly. Pro includes twice Starter’s usage and Slack/Jira integrations. Enterprise is quoted separately. These are the site’s EUR amounts, not local-currency offers; taxes may apply. Unlimited repositories does not establish unlimited reviews or tokens: confirm the actual allowance before buying.
Where your source code and review data go
The privacy policy says raw source code and diffs are retrieved for analysis rather than stored at rest in Revix’s database. Relevant code is nevertheless sent to AI providers including OpenAI and Anthropic. Revix states that this API-submitted data is not used to train their models.
Review findings, summaries and derived repository knowledge are retained while the workspace and repository remain connected; deletion can be requested. Review the data flow before granting access to confidential repositories. Your organization keeps its source-code rights, and AI findings can be inaccurate or incomplete.
Frequently asked questions
Can Revix review AI-generated code?
Yes. Revix applies the same PR review workflow whether code was written by a developer or an AI assistant. Findings still need validation.
Is Revix free for private repositories?
The permanent free tier is for public repositories. Private repositories are covered by the trial and then paid plans.
Does Revix replace a linter or human reviewer?
It adds repository-context checks alongside those tools. Its terms do not guarantee defect detection, so retain linting, tests and human judgment.
Is source code used to train models?
Revix says the API providers do not train on submitted code. That does not mean all processing stays local or that generated review data is immediately deleted.
Official resources
- Official setup and review examples: see the supported workflow.
- GitHub installation and permissions: inspect app access before connecting.
- GitLab connection guide: check roles and account attribution.
- Official team and contact information: use the published team contact for support.





