Revue
revue.sh
- Free plan
- Windows
- Linux
- Web
- API
What Revue is
Revue is a multi-agent AI code reviewer for local changes and pull requests. Its Claude Code skill reviews staged changes, while its CI mode can review pull requests or merge requests on GitHub, GitLab, and Bitbucket and post inline comments. Six specialist agents cover security, performance, code quality, architecture, synthesis, and verification. Revue says it checks findings against the code and anchors them to relevant lines. The FAQ lists OpenRouter, Anthropic, OpenAI, Azure OpenAI, and custom OpenAI-compatible endpoints as supported AI providers. Users must supply their own provider API key; Revue does not provide a shared AI backend. For local reviews, the privacy policy says code runs in the user's environment and source code is not received, stored, or processed by Revue's servers. The default diff limit is 2,000 lines and can be raised in .revue.yml. Windows use requires WSL2; there is no native Windows build, and Windows on ARM is unsupported. Free includes unlimited reviews for 30 days, then 10 per month. Indie costs $9 a month for 100 monthly reviews, and Pro costs $29 a month for unlimited reviews. A 30-day trial is offered.
Who it suits
Revue may suit developers who want local reviews before committing or teams seeking automatic reviews in pull-request workflows. Users need to bring an API key from a supported AI provider.
Good to know, and what to watch
- Six agents cover security, performance, quality, and architecture.
- Supports GitHub, GitLab, and Bitbucket pull requests.
- Local review source code stays in the user's environment.
- Supports custom review rules and review analytics.
- Pro includes unlimited monthly reviews.
- Users must supply their own AI provider API key.
- Free drops to 10 reviews per month after 30 days.
- Windows requires WSL2; Windows on ARM is unsupported.
- The default diff limit is 2,000 lines.
Plans
Every plan and priceFree
Indie
Pro
Free plan · 30-day trial
The verdict
Revue combines local and CI review workflows with multiple specialist agents and anchored findings. Check its API-key requirement, Windows constraints, and changing Free allowance against your workflow.
What the maker publishes
- AI providers
- The FAQ lists OpenRouter, Anthropic, OpenAI, Azure OpenAI, and custom OpenAI-compatible endpoints as supported providers. source
- AI providers
- Supported provider choices include OpenRouter, Anthropic, OpenAI, Azure OpenAI, and custom OpenAI-compatible endpoints. source
- Bring your own key
- Revue requires users to provide their own AI provider API key and says it does not provide a shared AI backend. source
- BYOK
- Users must supply their own AI-provider API key and therefore control provider selection and inference spending. source
- CI installation
- CI setup installs the revue-ci Python package and uses repository secrets or variables for license and AI keys. source
- CI workflow
- Revue can automatically review pull requests or merge requests in GitHub Actions, GitLab CI/CD, and Bitbucket Pipelines. source
- Cloud data processors
- The privacy policy names Stripe, Brevo, Cloudflare, and Fly.io as service subprocessors. source
- Code handling
- For local reviews, the privacy policy says code review runs in the user's environment and source code is not received, stored, or processed by Revue's servers. source
- Configuration depth
- The configuration reference documents provider, model, endpoint, temperature, token limits, review behavior, noise filters, agents, and output settings. source
- Contradiction detection
- A synthesis layer is intended to resolve disagreements between specialist agents. source
- Custom rules
- Repositories can define custom review behavior in a .revue.yml file. source
- Data collected
- The privacy policy says Revue may log code-change metadata, review counts, timestamps, repository and platform metadata, and diagnostic data. source
- Data handling
- For local execution, Revue says source code stays on the user's machine and only activation metadata and aggregate usage counts reach Revue servers. source
- Diff limit
- The default diff limit is 2,000 lines; the FAQ says it can be raised in .revue.yml. source
- Diff limit
- The default review limit is 2,000 unified-diff lines, adjustable through .revue.yml. source
- Fix suggestions
- Paid tiers can post platform-native code suggestions when the resolver generates a sufficiently confident scoped fix. source
- Free quota behavior
- After the initial unlimited period, the Free tier allows 10 reviews per calendar month and returns a 429 response after the limit. source
- Independent identity warning
- Crunchbase's Revue listing describes a different Utrecht newsletter company, Revue Holding B.V., acquired by X; it should not be treated as the revue.sh maker. source
- Independent identity warning
- The Chrome Web Store listing named RevueAI is a separate GitHub PR review extension by Billmike, not the revue.sh service. source
- Independent repository status
- The separate Billmike/revue-ai GitHub repository describes a Chrome extension and shows 1 star and 0 forks. source
- Intended users
- The site describes local reviews before committing and CI reviews for teams that want pull requests reviewed automatically. source
- Intent drift
- Revue compares acceptance criteria in a pull-request description with the changed code and reports per-criterion alignment or contradiction. source
- Licence model
- The terms grant a limited, non-exclusive, non-transferable internal-use licence and state that Revue retains ownership of the service and findings. source
- Local workflow
- Users install the /revue Claude Code skill and run reviews on staged diffs before committing. source
Checked against the maker's pages on 21 Oct 2026.