CodeRabbit - Self-Hosted AI Code Reviews

What do you like best about the product?

The speed and accuracy of the automated PR summaries and walk-throughs are incredible. It perfectly understands the context of complex code changes and provides immediate, actionable feedback, which significantly reduces the cognitive load on human reviewers.

What do you dislike about the product?

It occasionally leaves a high volume of comments on minor code refactoring suggestions or styling choices that don't match our exact internal preferences. It would be helpful to have slightly more granular control over the threshold for what it considers a critical note versus a minor nitpick.

What problems is the product solving and how is that benefiting you?

CodeRabbit helps remove the bottleneck of manual code reviews and the tedious work of hunting down hidden bugs in complex scripts. Rather than spending hours tracing logic paths or trying to catch minor syntax issues in automated workflows, CodeRabbit surfaces them immediately right inside the pull request. For me, that means faster deployments, more confidence that scripts are robust before execution, and consistently high code quality without slowing down development.

What do you like best about the product?

It works well for reviewing individual commits and similar changes.

What do you dislike about the product?

It doesn’t work well with large commits or bigger projects. It just keeps running indefinitely and eventually freezes.

What problems is the product solving and how is that benefiting you?

I’ve created agents that can take on the work for me. If you’re not very familiar with architecture or building things, it can be a great help and make the process much easier.

What do you like best about the product?

The feedback is good and has helped identify issues with code before going into testing, giving me more confidence prior to merging.

What do you dislike about the product?

The fair-usage is awful, and I'm not saying that because the licensing model isn't fair, it's that they give you zero visibility in what you've actually used. The fair usage states "X amount of PR reviews" on your plan, but they don't tell you how many you've used, when your 7 days renews, or anything. It all happens in a black box.

I reached out to support, and they can't even tell me, you're just supposed to trust them (I'm throttled to 1 an hour, which happens when you've done 90+ reviews in the last 7 days). Totally fair.

What's not fair is that equates to 35% of my LIFETIME PRs, which I definitely didn't do in only the last 7 days.

I also understand that a PR might be reviewed more than once. Fine. However, don't tell us to just "believe you" and the number is right, without showing us how it's actually calculated.

This lack of transparency makes us go a different direction when I'd easily spend more on the product if I felt the numbers were right because you've proved they are.

What problems is the product solving and how is that benefiting you?

Giving a second opinion when developing with Claude Code.

What do you like best about the product?

I like how CodeRabbit analyzes all the lines of change and adds relevant comments, finding potential issues. This feature is particularly valuable because it sometimes catches loopholes that a manual reviewer might miss.

What do you dislike about the product?

Good overall, but sometimes it adds irrelevant comments which can be skipped with broader context. Also, it adds too many things sometimes.

What problems is the product solving and how is that benefiting you?

I use CodeRabbit for reviewing pull requests. It analyzes all lines of change, adds relevant comments, and finds potential issues, sometimes catching loopholes that a manual reviewer might miss.

What do you like best about the product?

It's ability to review changes and understand them within the context of the work that is being done (along with the PR summary), and then to give very clear, structure feedback, with code diffs. It integrates very well with our current workflow, is affordable, and can quickly produce prompts for my AI for handover.

What do you dislike about the product?

Very occasionally it can make the wrong assumptions, but it does learn from the feedback it receives.

What problems is the product solving and how is that benefiting you?

It helps provide invaluable PR review support in a very busy environment where time is a major factor.

What do you like best about the product?

I really appreciate how CodeRabbit significantly reduces the reliance on another developer in the code review process, allowing me to continue my work in minimal time. It gives me the confidence that my code does not include serious bugs and code smells, which is incredibly reassuring. I also enjoy its seamless integration with GitHub Actions, making it easier to respond to comments directly. The initial setup of CodeRabbit was very easy, which saved me a lot of time and effort.

What do you dislike about the product?

I find it problematic that, like other AI tools, sometimes CodeRabbit becomes unstoppable and generates useless comments. This can be frustrating and require additional effort to handle.

What problems is the product solving and how is that benefiting you?

I use CodeRabbit to streamline code reviews, removing the bottleneck of waiting for another developer's input. It increases my confidence by identifying bugs and integrates seamlessly with GitHub Actions.

What do you like best about the product?

The product itself has proven quite useful. It has already spotted a great number of issues that we definitely would not have spotted ourselves. We rely on it every single day.

It's pretty easy to get started and to customise the rules and settings on the online panel - although jumping between repo settings and org settings is a bit awkward UX-wise.

The sales and onboarding processes were very accommodating, even if a bit slow.

What do you dislike about the product?

By far the biggest downside of CodeRabbit is their customer support. They have a chatbot that only exists to pre-fill an email.

Despite the bot asking for my email address (which they already have on file), they sent the response to my request to our billing contact's email instead. When I pointed this out as a fairly glaring security lapse, their response completely ignored that. Further contacts went unanswered entirely.

What problems is the product solving and how is that benefiting you?

CodeRabbit's automated PR reviews are pointing out edge cases and minor security issues in our code that human reviewers would've been unlikely to spot. The additional fortification increases our code quality and frees up human capacity for higher value engineering tasks.

What do you like best about the product?

It's pretty good to maintain good code quality and prevent potential bugs, it catches them directly in the PR and even suggest code changes directly, saves tons of time. In case of false positive, you can easily tell it to ignore it next time and it'll keep it in mind for future PRs, same for code style, preferences, etc.. Pretty much anything

What do you dislike about the product?

Although it is pretty good and I'm 99% happy with what it suggests, it can happen that some times some suggestions arent that great or valuable, but this is an AI and it's pretty much to be expected, you can always easily discard them and let it know so it doesn't do it again.

What problems is the product solving and how is that benefiting you?

I'm pretty much the only person working on a project and it makes it so I have "someone" else watching over the code I make

What do you like best about the product?

- easy to use, easy to converse with and interact with
- easy to implement

What do you dislike about the product?

I wish there was a progress meter or something when it is reviewing.

What problems is the product solving and how is that benefiting you?

- right now we have a small engineering team, and CodeRabbit gives us the extra security in having PRs reviewed, when we have limited engineers to run reviews currently.

What do you like best about the product?

- It explains analyzed PRs with diagrams and detailed descriptions, which really helps to review them later and make sure that the code does exactly what was expected
- It provides good quality code reviews, detecting bugs, not optimal implementations, missing tests, and suggests improvements
- It learns from feedback and communication with humans and does next reviews better
- It saves PR reviewers a lot of time by checking all the prerequisites.

What do you dislike about the product?

- It is unstoppable in its suggestions, providing comments and change requests even to the code, it suggested in previous iterations, so the process can run forever
- It still makes mistakes, and even after I ask to verify the suggestion or the fix, it is going to post, before the posting, it still doesn't do that, so we need to run another iteration of our discussion to verify it and correct if needed.

What problems is the product solving and how is that benefiting you?

We have many PRs from our community and lack of time fro review them by our maintainers of the project. CodeRabbitAI adds PR summary and explains the implementation. Also, it does code review before our maintainers and authors of the PRs can fix found issues even before a maintainer spends time on the review.