CodeRabbit

Introducing Project Atlas 🌐 The first AI-native code review interface. Try it in your PR's today!

Who's blocked? Which PRs are rotting? Why? CodeRabbit Agent for Slack tells you. Every hour (if you want it to).

@liran_tal Well

best code review agent?

@yoimkonrad They just don't understand

Oh did I do this video? Why don’t you have CodeRabbit Agent for Slack in your workspace yet?

Well shit, this wasn't on my 2026 bingo card but here we are. Thanks @coderabbitai for the sponsorship! Logo is now up in a brand new section on github.com/agavra/tuicr and tuicr.dev

@almoggavra Pleasure to sponsor you, tuicr is awesome 😌

@coderabbitai oops 🙊

Stop reverse-engineering PRs to review them. CodeRabbit walkthroughs now explain how the change was built and route it to the right reviewers. coderabbit.link/explainable-pr…

@buildwtim Time to install CodeRabbit and touch some grass Time 😅

@coderabbitai f this works, my coffee budget is about to take a serious hit. maybe i'll finally see daylight ;)

Cut Code Review Time and Bugs in Half. Sign up for a free trial!

Is Google scared of open-source models? @kunalstwt made a good case for why open source AI keeps gaining ground. We talked about Kubernetes, cloud automation, GPU shortages, and why engineers are tired of babysitting clusters. New episode of @coderabbitai's podcast, The Merge with Kunal from Cast AI.

Do you want to review a 1-million-line PR? @jarredsumner did not make it easy for either you or GitHub. Well, Atlas did! You can check how the agent went on to write the code on the left, and understand the line-block changes on the right. Comments in a separate tab.

@isaac_ts_way @aravindputrevu

@aravindputrevu @coderabbitai Awesome I will dig into this more and set it up.

All AI code review products suck ass right now. This is what I need: 1. Let me setup multiple parallel agents that look for different things. Multiple agents each doing one thing perform much much better than one omni review agent tasked with 14 things. Everyone knows this! 2. Let me have agent specific settings that let me specify how their feedback should be handled. There should be a “flag for human review” option, that requires a human to look at it. There should be an “automatically fix with model” option, that allows me to have any model I want fix it. This would allow me to: 1. Setup an “access control review” agent. Any access control issues get flagged for human review. 2. Setup a “query optimizer” agent, that makes sure all database queries are efficient, no n+1, batch writes, indexes, etc, and have agents auto fix them. (AI is really bad at just outputting non terrible database operations). And much more. Current AI review tools have not seen serious improvement since their original creation. Where is the iteration? Is there something like this already? @coderabbitai is still really useful but just could be so much more!

@usememos 🧡🧡

CodeRabbit has been reviewing our PRs, and now they’re sponsoring Memos too. Thanks @coderabbitai for both the code review nits and the open-source love! coderabbit.ai

you reviewed it lol

@ptr_to_joel 😮‍💨

holy wow they merged it

+1,009,257 -4,024 Honestly don't know how we ended up here...