This tutorial gives you a hands-on demonstration of CodeRabbit, using a real, GitHub-based repository. It guides you through the following tasks:

1. Integrate CodeRabbit into a GitHub-based repository that you own.
2. Observe CodeRabbit perform a code review of a pull request that you initiate.
3. Converse with CodeRabbit about the code review.
4. Prompt CodeRabbit to generate its own improvements to the pull request.

When you complete this tutorial, you'll have seen CodeRabbit's code-review feature in action, and glimpsed a few of its other AI-driven abilities as well.

For a more general overview of CodeRabbit, seeIntroduction.

Before you begin​

Create a new, private repository on GitHub. Name the new repositorycoderabbit-test, and let it have otherwise default GitHub settings.

Integrate CodeRabbit with your GitHub account​

To integrate CodeRabbit with your GitHub account, follow these steps:

1. Visitthe CodeRabbit login page.
2. ClickLogin with GitHub.
3. ClickAuthorize coderabbitai.

CodeRabbit takes a moment to set up the integration. After it finishes, the CodeRabbit dashboard appears.

Add CodeRabbit to your repository​

To add CodeRabbit to your test repository, follow these steps:

1. On the CodeRabbit dashboard, clickAdd Repositories. A GitHub repository-access dialog appears.
2. Select theOnly select repositories radio button.
3. From theSelect repositories menu, select thecoderabbit-test repository that you created earlier in this Quickstart.
4. ClickInstall & Authorize.
   note

   CodeRabbit requests read and write access to your repository in order for its code review, issue management, and pull request generation features to work. CodeRabbit never stores your code. For more information, seethe CodeRabbit Trust Center.

5. If a CodeRabbitComplete your signup dialog appears, then fill it out with the requested information before continuing.

CodeRabbit is now ready to use with your test repository. The next steps demonstrate its core code-review features.

Let CodeRabbit perform a code review​

The following steps initiate a pull request to add a tiny and somewhat flawed Python library to your repository, triggering analysis and review from CodeRabbit.

Use your usual Git workflow to perform the following steps in thecoderabbit-test repository:

1. Create a branch namedadd-utils.

2. In that newadd-utils branch, create a new file calledsimple_utils.py, with the following content:

   # simple_utils.py - A tiny utility library
   
   defreverse_string(text):
   """Reverses the characters in a string."""
   return text[::-1]
   
   defcount_words(sentence):
   returnlen(sentence.split())
   
   defcelsius_to_fahrenheit(celsius):
   return(celsius*9/5)+32
   

3. Commit the added file to theadd-utils branch. Use any text you want for the commit message.

4. Create a pull request that proposes to merge theadd-utils branch into themain branch. Use any text you want for the pull request message.

After a few moments, CodeRabbit responds to the pull request using the@coderabbitai GitHub account. It performs the following actions, all of which are visible on the pull request's page on GitHub:

• If you didn't write a pull request summary, then CodeRabbit adds a new summary to the pull request.
• CodeRabbit posts a comment titledWalkthrough containing analysis and commentary about the content of the pull request.
• CodeRabbit attaches a detailed code review to the pull request as another comment.

Part of the review resembles the following screenshot:

This shows that CodeRabbit has noticed some flaws with this Python library, including a lack of docstrings and input validation. The review comment identifies these flaws, and suggests how you might improve them.

Ask CodeRabbit for more information​

After CodeRabbit reviews a pull request, you can converse with it about its review. You can also give CodeRabbit special commands to request it to perform specific functions, such as generating appropriate code changes for you.

1. Post the following as a new GitHub comment attached to the pull request:

   @coderabbitai Why do all of these functions need docstrings? Isn't it obvious enough what they do?
   

   After a few moments, CodeRabbit posts its reply, explaining its motivations for its recommendation. The reply might resemble the following:

   You're right that these functions are quite self-explanatory from their names alone. Docstrings aren't strictly necessary for every function, especially in smaller projects with clear function names.

   While docstrings do provide benefits (IDE tooltips, documentation generation, help for future developers), they can sometimes feel redundant for very straightforward functions. It's ultimately a matter of coding style and project requirements.

   If you prefer to skip docstrings for obvious utility functions in this project, that's perfectly reasonable. Would you like me to adjust my review approach regarding documentation for this repository?

   In this case, CodeRabbit is offering to adjust its future behavior when reviewing changes to this repository. Instead, follow the next step to ask CodeRabbit to implement part of its suggestions.

2. Post the following as another new comment:

   @coderabbitai generate docstrings
   

   After a few moments, CodeRabbit does the following:

   • CodeRabbit creates a new branch, based onadd-utils.
   • CodeRabbit commits changes to the branch that adds suggested docstrings toadd-utils.
   • CodeRabbit opens a new pull request between the new branch andadd-utils.

Clean up​

Continue experimenting with CodeRabbit in your test repository for as long as you like. When you're finished, you can delete the repository.

You are now ready to add CodeRabbit to other repositories that you own, and let it review some real pull requests.

What's next​

• Integrate CodeRabbit with your repositories on GitHub, GitLab, Azure DevOps, or Bitbucket.
• Configure CodeRabbit beyond its default settings.
• Add custom review instructions.
• Get support for CodeRabbit.
• Learn more about how CodeRabbit works.