We're excited to announce the addition of two powerful tools to our static analysis arsenal:

- **Brakeman**: A static analysis security vulnerability scanner for Ruby on Rails applications. It helps identify security issues in your Ruby codebase by analyzing Gemfile, Ruby files (_.rb), and ERB templates (_.erb).

- **Clippy**: The official linter for Rust code, helping catch common mistakes and improve your Rust code quality. It analyzes \*.rs files and supports configuration through clippy.toml files.

Both tools can be configured through their respective config files or through CodeRabbit's settings page. See our [tools documentation](https://docs.coderabbit.ai/tools/) for more details.

We're thrilled to announce that CodeRabbit's AI code reviews have officially landed inside **VS Code and all its forks, including Cursor and Windsurf!** No more review bottlenecks or context switches: now you can get _instant, inline feedback right in your editor, completely free._

**Why you'll love it:**

- 📝 **Inline Code Reviews**: Senior-dev-level annotation on every line. CodeRabbit becomes your AI pair programmer, annotating staged & unstaged commits before you even open a PR.

- 🚀 **Built for Flow, Not Friction**: Code, review, commit—rise & repeat without breaking your flow state. Reviews run automatically on every commit so you ship faster.

- 🛠️ **Fix-with-AI**: One-Click Fix for quick tweaks; "Fix with AI" hands off more complex suggestions to your favorite coding agent, complete with full context.

- 🧩 **Compatible with Cursor, Windsurf & Language-Agnostic**: Works out of the box on VS Code, Cursor, Windsurf. Supports Java, JavaScript, PHP, Python, TypeScript, Go, Ruby, and more.

Pair these IDE-embedded reviews with our deeper Git-platform reviews for a multi-layered approach that catches more bugs and cuts down overall review time.

- [Join the launch discussion on Discord](https://discord.com/channels/1134356397673414807/1145813948193575023/1372210767838842950)

- [Install the CodeRabbit VSCode Extension](https://coderabbit.link/vscode)

| Reviews per hour | 3 back-to-back, then 2/hour (Summary only) | 4 back-to-back, then 3/hour | 3 back-to-back, then 2/hour | 5 back-to-back, then 4/hour |

title: Code review best practices

description: Best practices for managing CodeRabbit code reviews.

sidebar_label: Best practices

This page lists best practices for performing code reviews with CodeRabbit.

For more information about working with

CodeRabbit through chat, see [Control and manage code reviews](/guides/commands).

For a CodeRabbit command reference, see [Code review command reference](/reference/review-commands).

- Start with `@coderabbitai review` for checking new changes.

- Use `@coderabbitai full review` when major changes require a fresh perspective.

- Generate summaries after significant updates using `@coderabbitai summary`.

- Use `@coderabbitai pause` before making multiple commits.

- Resume reviews with `@coderabbitai resume` when ready.

- Consider `@coderabbitai full review` after substantial changes.

- Run `@coderabbitai generate docstrings` after finalizing function implementations.

- Learn more about [docstring generation](/finishing-touches/docstrings).

- Commands are case-insensitive (`@coderabbitai REVIEW` works the same as `@coderabbitai review`).

- Commands can be issued by anyone with write access to the repository.

- Multiple commands can be used in sequence as needed.

- Use `@coderabbitai configuration` to export your settings before making changes.

- Most commands (pause, resume, ignore) take effect immediately.

- Review commands typically complete within a few minutes, depending on PR size.

- Docstring generation time varies based on the number of functions.

title: Review pull requests

description: An overview of CodeRabbit's core code review features.

sidebar_label: Overview

The central feature of CodeRabbit is its ability to proactively review

new pull requests on your code repository.

CodeRabbit reviews take the form of pull request comments that

include summaries, analyses, and initial critiques of the proposed changes.

This information, usually added to pull requests within minutes, can help your team perform more rapid, better-informed code reviews.

The following sections present an overview of this feature. For a hands-on example that lets you experience a CodeRabbit code review using

a real repository, see [Quickstart](/getting-started/quickstart).

After you [integrate CodeRabbit with your repository](/platforms), CodeRabbit proceeds

to automatically review every subsequent pull request, as soon as each one is created.

CodeRabbit performs code reviews by attaching comments to the pull request.

These comments contain detailed summaries and analyses of the changes,

as well as listing out problems or areas for potential improvement that it found.

CodeRabbit uses [a variety of open-source linters and security tools](/tools) and a custom

code verification agent to provide this analysis. CodeRabbit

also consults several models to further analyze and critique the proposed changes,

using all of the content of your repository as context. The code-review comment that CodeRabbit attaches

to your pull request synthesizes and summarizes all of the information collected from these different sources.

For more information about the graph analysis that CodeRabbit includes with its reviews

when available, see [CodeRabbit Code Graph Analysis](/integrations/code-graph-analysis).

By default, the following activity in your repository triggers CodeRabbit to

perform a code review:

- If CodeRabbit sees a new pull request, then it immediately performs a full review

of the proposed code changes.

- If an open pull request that CodeRabbit has already reviewed gets modified with another

commit, then CodeRabbit performs an incremental review that focuses on the new commit.

After CodeRabbit attaches its initial code-review comment to a pull request, you can

directly interact with CodeRabbit by mentioning its username, `@coderabbitai`, in comments

that you post to the pull request.

These interactions can serve several purposes:

- Free-form discussion about the pull request and the ongoing code review.

- Commands to have CodeRabbit perform specific actions regarding the code review.

- Prompts to have CodeRabbit generate its own improvements to the branch under review.

You can have open-ended, natural-language discussion with CodeRabbit during a code review, treating it

as an LLM-powered chatbot that has your entire code repository available for context. For more information, see [CodeRabbit Chat](/guides/agent_chat).

CodeRabbit recognizes a variety of keyword-based commands that let you control its

behavior during a code review, including the following:

- Pause or resume automated reviews of the pull request.

- Manually request a review, when automated reviews are paused.

- Resolve all open comments authored by CodeRabbit.

For more information, see [Control and manage code reviews](/guides/commands).

You can command CodeRabbit to generate improvements to the branch under review.

CodeRabbit accomplishes this by publishing a new branch based on the branch under review,

and creating a new pull request for your own review.

Available code-generation commands let you request the following from CodeRabbit:

- Implement the suggestions for improvements that CodeRabbit has made in its earlier code review comments.

- Generate inline documentation for any undocumented functions that this pull request proposes to add.

For more information, see [Generate improvements](/guides/generate-improvements).

- [Control and manage code reviews](/guides/commands)

- [Generate code improvements](/guides/generate-improvements)

title: Code review troubleshooting

description: Troubleshooting CodeRabbit code reviews.

sidebar_label: Troubleshooting

This page is about troubleshooting interactive code review sessions with CodeRabbit. For more information about working with

CodeRabbit through chat, see [Control and manage code reviews](/guides/commands).

If a CodeRabbit command doesn't seem to work:

1. Check that you have the necessary repository permissions.

2. Verify the command syntax.

3. Look for any response from CodeRabbit in the PR comments.

4. Use `@coderabbitai help` for command guidance.

Need help? Join our community on [Discord](https://discord.gg/coderabbit) or [contact our support team](/getting-started/support).