Subscribe
Sign in
CodeRabbit — Cut code review time & bugs in half using AI
Get automated, AI-powered code reviews that surpass existing tools. It offers real-time collaboration, fine-tunes reviews based on user feedback, and uses a configurable rule engine.
Eliminate bugs, ship fast, and let developers focus on big problems.
45
Replies
Best
@vishavjeet_kaur Very cool. Reducing some of the busywork in engineering will always get my upvote.
Can you explain more about its retrieval strategies? I assume it pulls the repository to reference code from source files that aren't present in the PR.
@vishavjeet_kaur Does this capture the potential credentials developers accidentally commit to github? I can totally see this as the next level cyber security tool!
@vishavjeet_kaur @sixbangs That's correct. We clone the repo in a secure sandboxed environment.
Congrats on the launch @vishavjeet_kaur 🚀
What does CodeRabbit do when it 'validates PR changes with Linear'?
@vishavjeet_kaur @jgani.
We validate the PR against the linked issues (Linear, Jira, GitHub, GitLab issues etc) and post the result on the PR.
We have been trialing it and I like it overall but when it came time to decide to pay, we’re holding off for now. Here’s some hopefully helpful feedback:
1. It’s too noisy. Lots of positive affirmations about our code makes it hard to find the actual meat of the review. There’s lots of comments about the status of the review which end up clogging my inbox.
2. It goes back and reviews everything over again (at least that’s how it seems) if I push a small commit on top of a PR. For example if I push my code up, and it reviews the code, then I push up a spelling change, it kicks everything off again.
Overall lots of promise and wishing you good luck. We’ll probably be back in the future!
@willewallace I have had the same experience. The noise is also in terms of the actual review as well.
The bot is sometimes not able to see how errors are handled either in the caller or the callee functions and keeps adding a review comment which just has to be marked as resolved.
My team has stopped taking its comments seriously.
Our tool is open source, so we can use CodeRabbit for free. I am still wondering if we should continue to use it or not because sometimes human review comments get lost in the noise of CodeRabbit review comments.
@vishavjeet_kaur I didn't see the bottom price there 😂 that's hard to spot. I never scrolled that far. 😅 But that's amazing! will try tonight then 🎉
Code Rabbit is an impressive tool that helps OSS software to improve their code review. CodeRabbit is different from other superficial AI product: it uses plenty of retrieval strategies to extract information from the codebase. As a result, the rabbit's review comments are usually insightful and make you feel they actually understand the code.
Definitely recommended for both team devs and solo devs.
Amazing tool, congrats @gursingh @vishavjeet_kaur !
Am interested to know more about # 4 Auto-generation of learnings from the user feedback - is there an integration involved here?
Launching soon!
Congrats on the launch, @vishavjeet_kaur !
We've tried CodeRabbit for jethire.ai and the team feedback was mostly positive. Well, to be honest, most of the time we were impressed how well it catches all the details in quite large and complex codebase. There were some minor issues, though, but they are outweighed by the positive side.
I was closely working with AI/ML for a few years, and obviously tried to build something similar in spare time and talked a lot with people who tried to make more or less similar product, however we struggled with navigating large code base and extracting relevant parts of the code to be analysed. The idea how you used "scripts" for code analysis is really fascinating.
The issues we've noticed are purely false positives, i.e. it suggests reformatting part of the TS code in a way that contradicts with our code style.
Will we use CodeRabbit after trial period? Maybe. It is quite pricey for bootstrapped startups (and solopreneurs); however, I believe, it has a large market.
Great to see something like this. It's actually really easy to get locked into the same things to check for in a code review over and over and just get stuck on autopilot so it will be great to use this and get out of that same zone. The configuration on it is awesome.
@dsned This seems to be your first PH post.
Thank you for taking the time to create an account today to post about our launch.
Would love to get on a call to understand your perspective. Would you be open to it?
Meanwhile, feel free to browse through 100k+ CodeRabbit reviews on public repositories.
https://github.com/search?q=code...
CodeRabbit is a game-changer for the Wechaty community! It dramatically speeds up PR reviews with AI-powered insights that feel like they're from an expert developer. Its real-time collaboration and configurable rule engine make it stand out, helping us ship faster and with fewer bugs. Highly recommend for any team aiming to improve their code quality and efficiency! 🚀🐰
🎉Congrats! We're launching "Cascad3 - A decentralized inspiration-driven publishing platform" on Product Hunt soon. Let's collaborate! I would appreciate your support and feedback by clicking the "Notify Me" button to stay tuned and be the first to know when Cascad3 debuts on the stage.
👇
https://www.producthunt.com/prod...
Thank you!
Best regards,
Yami
Congrats on the launch, CodeRabbit seems like a great tool! Can you share insights on how the AI adapts to different coding styles and languages to ensure accurate and relevant feedback?
@vasily_kolosov1 Reviews are fine-tuned based on the auto-generated learnings from human feedback, configurable natural language instructions, and a rule engine that allows setting coding standards using an intuitive pattern language.
Congratulations on the progress and success of CodeRabbit 🤍 It's fantastic to hear about the new features and enhancements you've introduced to benefit the developer community. The addition of comprehensive code verification, integration with Linear & Jira, and the auto-generation of learnings from user feedback are particularly impressive.
How does CodeRabbit ensure the accuracy and reliability of its contextual daily standup reports?
@yasec_kowalski Thanks! Reports are generated based on the pull request activity during the day, with code as the source of truth. We summarize the code changes to build these reports.
I can already envision how this tool will change the way we approach code quality and release cycles. With automated code reviews and bug detection, we can ship faster and with more confidence. And with developers freed up from tedious manual reviews, we can focus our efforts on tackling bigger, more complex problems that truly drive innovation.
CodeRabbit