Automated AI code review on every pull request
Install a GitHub App. Bring your Anthropic, OpenAI, or Gemini key — or an OpenRouter key for open-weight models like Kimi, GLM, DeepSeek, and Qwen. Get detailed, context-aware code reviews — automatically, in minutes.
14-day free trial. $10/mo. No credit card required.
How it works
- Install the GitHub App — Add CodeMouse to your GitHub organization or personal account. Choose which repos to connect.
- Add your AI key(s) — Connect Anthropic, OpenAI, Gemini, or OpenRouter (Kimi, GLM, DeepSeek, Qwen, and 100+ more). Connect several and each PR gets reviewed by every model in parallel. Keys are encrypted at rest (AES-256-GCM).
- Open a pull request — CodeMouse reviews it automatically and posts detailed comments directly on GitHub.
Built for developers
- Full repo context — Reviews use the entire repository, not just the diff.
- Claude, GPT, or both — Connect one or both. When both are connected we merge the findings and drop duplicates, so you get a single consolidated review.
- Bring your own key — No AI cost markup. You pay your provider directly (~$0.05–0.15 per review per provider; open-weight models via OpenRouter cost a fraction of that).
- PR commands — Comment
@codemouse review on any pull request to trigger a review manually (no push needed), or @codemouse full review to re-review the entire diff. Follow-up pushes get incremental reviews automatically.
- Reads the room — Reviews every existing PR comment (humans, other bots, earlier CodeMouse runs) and won't repeat what's already been said. Follows up on its own threads when fixes land.
- Approves clean PRs — When the diff looks good, prior concerns are addressed, and CI is green, CodeMouse actually approves the PR.
- No throttling — Every review starts instantly. No queues, no rate limits.
- Works with any language — TypeScript, Python, Go, Rust, Java — if it's code, CodeMouse reviews it.
- Inline comments — Review comments appear on specific lines in your PR.
- Zero config — No CI pipelines, no YAML files, no build steps.
Simple pricing
$10/month with a 14-day free trial. Unlimited repositories, pull requests, and team members. No throttling — reviews start instantly. Plus your own AI API usage (~$0.05–0.15 per review per provider; less with open-weight models via OpenRouter).
Compare with other AI code review tools
See the head-to-head comparisons:
Frequently asked questions
- What AI models does CodeMouse use?
- Your choice: Anthropic Claude, OpenAI GPT, Google Gemini — plus open-weight models like Kimi, GLM, DeepSeek, and Qwen through OpenRouter. Connect several and each PR gets reviewed by every model in parallel; we merge the findings and drop duplicates.
- What happens when both models flag the same thing?
- We de-duplicate overlapping comments by file, line, and semantic similarity. You get a single consolidated review — not two bots arguing on your PR.
- Is my code stored?
- No. Your repository is cloned temporarily to run the review and deleted immediately after.
- Does it work with private repositories?
- Yes. The GitHub App uses installation access tokens scoped to the repositories you grant access to.
- How much does the AI usage cost?
- Typically $0.05–0.15 per review per provider (open-weight models via OpenRouter can be far cheaper). You pay your AI provider directly — we don't add any markup.
- Can I cancel anytime?
- Yes. Cancel through the billing portal. No contracts, no questions asked.
- Does it replace human code reviews?
- No — it complements them. CodeMouse catches mechanical issues so your team can focus on architecture and design.