CodeRabbit Signals

Counted for Engineering leaders and teams.

1. Positive

   ryanchants, Hacker News commenter

   Comment on: Ask HN: Does a human still review your code?Hacker News comment

   Says a human reviews their code less and less. Their employer runs CodeRabbit, which catches real issues, and they add their own checks on top.

   Published Aug 14, 2026self-described: Employer team that runs CodeRabbit on pull requests

2. Positive

   The_Fox, Hacker News commenter

   Comment on: There is an AI code review bubbleHacker News comment

   Installed CodeRabbit for their GitLab reviews; at roughly US$15 per user per month it regularly finds subtle problems, though it flags theoretical issues they just resolve.

   Published Jan 26, 2026self-described: Team using GitLab

3. Positive

   Cem Dilmegani, Principal Analyst at AIMultiple

   AI Code Review Tools BenchmarkIndependent benchmark, AIMultiple

   AIMultiple replayed 309 pull requests through CodeRabbit, Greptile, GitHub Copilot and Cursor Bugbot on default settings. CodeRabbit had the top LLM-judge score on 51% of them and also led the scores from 10 developers who graded a subset.

   "CodeRabbit ranked as the most successful code review tool across 51% of 309 PRs…"

   Published Jan 14, 2026self-described: AIMultiple's RevEval: 309 pull requests replayed on benchmark repositories for CodeRabbit, Greptile, GitHub Copilot and Cursor Bugbot, with 10 developers grading a subset

4. Positive

   Jörg Thalheim (Mic92), Nix team member

   Coderabbit github app for the Nix repoGitHub issue

   Requested CodeRabbit for the Nix repository as an opt-in reviewer after it found bugs in other projects the requester maintains. It later caught use-after-free bugs before they merged, while others objected to its poems and comment volume.

   Published Oct 27, 2025self-described: NixOS Nix repository, opt-in reviews per pull request

5. Positive

   Benoît Viguier (ildyria), Lychee maintainer

   AI-assisted reviews, one month later...Project blog post, Lychee blog

   Categorizing all 290 CodeRabbit findings on 28 PRs over a month, the maintainer rated 72% relevant and 28% noise or wrong assumptions, with real catches including Zip Slip and IDOR flaws.

   Published Sep 13, 2025self-described: LycheeOrg open source projects on CodeRabbit's open source plan

6. Positive

   Erin Doyle, founding engineer at Quotient

   The AI Revolution Fails without Psychological Safety for Developers: a Conversation with Erin DoylePodcast episode transcript, The InfoQ Podcast (The Architects Podcast)

   Quotient's founding engineer says the team uses CodeRabbit on its pull requests after trying several other tools. When faster AI coding left PRs stuck in review, the team agreed to leave bugs and edge cases to tools like CodeRabbit and keep human review for the business domain and maintainability.

   "We use CodeRabbit for our PRs. We've tried a number of different tools before landing on that one."

   "What did we think AI tools, such as CodeRabbit are really good at spotting and what do we feel like our responsibilities should be as human reviewers and what should we leave up to the AI?"

   Published Sep 28, 2026Quotient, a small engineering team building an engineering productivity product

7. Negative

   Martin Zoeller, Independent consultant who runs workshops on introducing AI into development teams

   CodeRabbit AI Review: Is It Worth It for Your Team?Independent review, MartinZoeller.com

   After weeks at a client, found reviews fast and some real frontend bugs caught. Noise, invented rules, wrong learnings and shallow context still made it not worth it next to a custom review skill.

   Published Jul 30, 2026self-described: Client engineering team, several weeks across dozens of pull requests

8. Mixed

   esposem, Agent Sandbox contributor

   [Discussion] Feedback for AI pre-reviewer (Coderabbit / Copilot) (kubernetes-sigs/agent-sandbox issue 1231)GitHub issue comment

   Finds CodeRabbit useful and mostly correct, but it drifts outside a pull request's scope, drops a finding as soon as a person disagrees, and can leave contributors talking only to a bot.

   Published Jul 21, 2026self-described: Kubernetes Agent Sandbox and other repositories that run CodeRabbit

9. Positive

   atestu, Hacker News commenter

   Comment on: Open Code Review, an AI-powered code review CLI toolHacker News comment

   Their company started using CodeRabbit in 2026. The automated review catches a lot, and the new review interface helps a person review large AI-written pull requests faster.

   Published Jun 5, 2026self-described: Company that adopted CodeRabbit in 2026 while trying to limit large AI-written pull requests

10. Positive

    Nick Tune, Senior Staff Software Engineer at PayFit

    Agentic code workflows with Nick TuneNewsletter interview, Tech World With Milan

    PayFit's senior staff engineer has CodeRabbit on every pull request at work, configured to read the team's coding standards and ADRs and flag violations, and sometimes runs the CLI before opening a branch. He reviews every PR himself after CodeRabbit and local review agents.

    "One thing that I use everywhere is CodeRabbit. It is quite amazing. I have it set up on all my pull requests, both personal projects and at work."

    "The great thing about CodeRabbit is that not only can it identify bad and insecure code, but it can also be configured to read all your coding standards and ADRs and find violations."

    "Sometimes I also run CodeRabbit via CLI before creating a branch."

    Published Mar 26, 2026PayFit, European payroll and HR company; Claude Code workflow with CodeRabbit on every pull request

11. Mixed

    Mark Di Marco, startup hacker from Austin

    Picking an AI Code ReviewerPersonal blog post, markmarkoh.com

    A five-person team trialed CodeRabbit, Cubic, Cursor's BugBot and Claude Code Review side by side. CodeRabbit caught real bugs, but its nitpicks and refactor suggestions meant the team would ignore 75% or more of its comments, so they chose Cubic.

    "While it caught real issues, it also has a lot of “Nitpicks” and “Refactor Suggestions”, so we’d ignore 75% or more of the comments it made."

    Published Oct 10, 2025self-described: Five-person startup engineering team trialing four AI reviewers with minimal customization

12. Positive

    Jonathan Petitcolas, Engineering Platform Lead at Tint

    The 3-5x Factor: Measuring the Real Impact of AI on Engineering TeamsCompany engineering blog post, Tint Technical Blog

    Tint's platform lead says AI coding raised pull request volume until review became the bottleneck, so the team brought back CodeRabbit, which had disappointed them a year earlier. Comments per PR rose 94% (1.61 to 3.12) after the July 3 rollout, which he counts as better review, and cycle times were trending back down.

    "…this spike is primarily due to implementing CodeRabbit on July 3rd. CodeRabbit dramatically increased the number of comments, but for good reasons: it catches issues that human reviewers might miss…"

    "This is precisely why we introduced CodeRabbit. The preliminary results are promising…"

    Published Sep 8, 2025Tint, an embedded insurance platform; the platform team measures engineering metrics before adopting tools

Counted for Engineering leaders and teams.

Positive, 4

1. Positive

   Maciej Korolik, Senior Frontend Developer and AI Expert at Monterail

   Best AI Code Review Tools 2026: Comparison & GuideCompany engineering blog post, Monterail blog

   Monterail ran a hands-on evaluation of GitHub Copilot, Cursor BugBot and CodeRabbit on an internal project and picked CodeRabbit, mainly for its PR summaries and sequence diagrams. Out of the box it posted many comments, and Monterail budgets one to two weeks of .coderabbit.yaml tuning.

   "At Monterail, we ran a formal hands-on evaluation of three leading tools: GitHub Copilot, Cursor BugBot, and CodeRabbit. CodeRabbit won."

   Published Oct 30, 2025Monterail, a software development company; GitHub Copilot, Cursor BugBot and CodeRabbit tested on one active internal project

2. Positive

   strawgate, FastMCP contributor

   Enable CodeRabbitGitHub issue

   Asked FastMCP to enable CodeRabbit after using it on another project: it writes useful pull request summaries and analyzes code well outside the changed files. The maintainer turned it on two days later.

   Published Oct 25, 2025self-described: PrefectHQ FastMCP open source project, after using CodeRabbit on py-key-value

3. Positive

   Benoît Viguier (ildyria), Lychee maintainer

   AI-assisted reviews, one month later...Project blog post, Lychee blog

   Categorizing all 290 CodeRabbit findings on 28 PRs over a month, the maintainer rated 72% relevant and 28% noise or wrong assumptions, with real catches including Zip Slip and IDOR flaws.

   Published Sep 13, 2025self-described: LycheeOrg open source projects on CodeRabbit's open source plan

4. Mixed

   Ahmed Atwa (nightknighto), AsyncAPI Generator contributor

   change coderabbit's config to be less annoyingGitHub issue

   Called CodeRabbit generally useful but too verbose and repetitive for small PRs; maintainers agreed to keep the summary and turn off the poem, suggested labels and reviewers, and make diagrams optional.

   Published Apr 24, 2025self-described: AsyncAPI Generator open source project, small frequent PRs

Mixed or negative, 1

1. Mixed

   Mark Di Marco, startup hacker from Austin

   Picking an AI Code ReviewerPersonal blog post, markmarkoh.com

   A five-person team trialed CodeRabbit, Cubic, Cursor's BugBot and Claude Code Review side by side. CodeRabbit caught real bugs, but its nitpicks and refactor suggestions meant the team would ignore 75% or more of its comments, so they chose Cubic.

   "While it caught real issues, it also has a lot of “Nitpicks” and “Refactor Suggestions”, so we’d ignore 75% or more of the comments it made."

   Published Oct 10, 2025self-described: Five-person startup engineering team trialing four AI reviewers with minimal customization

Counted for Engineering leaders and teams.

Positive, 4

1. Positive

   Janet Kuo (janetkuo), Agent Sandbox maintainer

   [Discussion] Feedback for AI pre-reviewer (Coderabbit / Copilot) (kubernetes-sigs/agent-sandbox issue 1231)GitHub issue comment

   Prefers CodeRabbit's review experience to Copilot's: it edits its existing comments without reposting them, is configurable, and learns when corrected, though Copilot catches some issues CodeRabbit misses.

   Published Jul 20, 2026self-described: Kubernetes Agent Sandbox project, running CodeRabbit and Copilot as pre-reviewers

2. Positive

   Nick Tune, Senior Staff Software Engineer at PayFit

   Agentic code workflows with Nick TuneNewsletter interview, Tech World With Milan

   PayFit's senior staff engineer has CodeRabbit on every pull request at work, configured to read the team's coding standards and ADRs and flag violations, and sometimes runs the CLI before opening a branch. He reviews every PR himself after CodeRabbit and local review agents.

   "One thing that I use everywhere is CodeRabbit. It is quite amazing. I have it set up on all my pull requests, both personal projects and at work."

   "The great thing about CodeRabbit is that not only can it identify bad and insecure code, but it can also be configured to read all your coding standards and ADRs and find violations."

   "Sometimes I also run CodeRabbit via CLI before creating a branch."

   Published Mar 26, 2026PayFit, European payroll and HR company; Claude Code workflow with CodeRabbit on every pull request

3. Mixed

   Federico Capoano (nemesifier), OpenWISP maintainer

   [ci] Opt in valid PRs to continuous CodeRabbit reviewsGitHub issue

   Because CodeRabbit reviewed every PR, including ones later flagged invalid, the maintainer turned off automatic reviews and had OpenWISP's bot request reviews only on valid PRs.

   Published Aug 31, 2026self-described: OpenWISP open source repositories

4. Mixed

   rparolin, NVIDIA cuda-python maintainer

   Enable CodeRabbit auto-review on cuda-python (port CCCL's .coderabbit.yaml)GitHub issue

   The cuda-python team had turned automatic review off because the default profile was noisy. It proposed turning it back on with the tuned config NVIDIA's CCCL team adopted to suppress walkthroughs, poems and finishing-touch suggestions.

   Published May 25, 2026self-described: NVIDIA cuda-python repository, CodeRabbit installed in opt-in mode

Mixed or negative, 4

1. Positive

   Michał Woźniak (mimowo), Kueue maintainer

   Feedback on CodeRabbit (kubernetes-sigs/kueue issue 13078)GitHub issue comment

   In Kueue's pilot review, calls CodeRabbit excellent and wants to keep it, but it forgets lessons the maintainers teach it in comments, such as filing flake issues.

   Published Jul 15, 2026self-described: Kubernetes Kueue project, a CodeRabbit pilot in kubernetes-sigs

2. Mixed

   stbenjam, Red Hat engineer

   CodeRabbit: Recommended setGitHub issue

   Asked Red Hat Product Security for a smaller must-have set of CodeRabbit checks. The product caps custom checks at 20, several duplicate what CodeRabbit or other scanners already do, and the AI-attribution check flags nearly every Claude pull request.

   Published May 26, 2026self-described: Red Hat Product Security's shared CodeRabbit configuration for other teams

3. Negative

   andrekandre, Hacker News commenter

   Comment on: Opus 4.5 is not the normal AI agent experience that I have had thus farHacker News comment

   Their team's CodeRabbit reviews seem to ignore the repository's CLAUDE.md conventions, and they are looking for a better reviewer.

   Published Jan 8, 2026self-described: Team using CodeRabbit with a CLAUDE.md file in the repository

4. Negative

   Martin Zoeller, Independent consultant who runs workshops on introducing AI into development teams

   CodeRabbit AI Review: Is It Worth It for Your Team?Independent review, MartinZoeller.com

   After weeks at a client, found reviews fast and some real frontend bugs caught. Noise, invented rules, wrong learnings and shallow context still made it not worth it next to a custom review skill.

   Published Jul 30, 2026self-described: Client engineering team, several weeks across dozens of pull requests

GILD's fact checks, which don't count as sources

Checked Oct 12, 2026.Learnings from review chat apply immediately by default. Set knowledge_base.learnings.approval_delay to 1 to 30 days to let an admin reject new Learnings first. Learnings

Counted for Engineering leaders and teams.

Raised as a concern, 18

1. Negative

   Kat Cosgrove (katcosgrove), Kubernetes contributor writing from the SIG Release perspective

   REQUEST: Enable CodeRabbit integration for kubernetes (kubernetes/org issue 6517)GitHub issue comment

   Opposes enabling CodeRabbit on kubernetes/kubernetes without a contract, a comparison with other tools including Copilot, and summon-only reviews, since many people find automatic reviews unhelpful.

   Published Aug 6, 2026self-described: Kubernetes project, discussing CodeRabbit for kubernetes/kubernetes

2. Negative

   Jordan Liggitt (liggitt), Kubernetes maintainer

   REQUEST: Enable CodeRabbit integration for kubernetes (kubernetes/org issue 6517)GitHub issue comment

   Says that on the pull requests they opened or followed, CodeRabbit's comments were only annoying and not helpful.

   Published Aug 5, 2026self-described: Kubernetes project, discussing CodeRabbit for kubernetes/kubernetes

3. Negative

   Martin Zoeller, Independent consultant who runs workshops on introducing AI into development teams

   CodeRabbit AI Review: Is It Worth It for Your Team?Independent review, MartinZoeller.com

   After weeks at a client, found reviews fast and some real frontend bugs caught. Noise, invented rules, wrong learnings and shallow context still made it not worth it next to a custom review skill.

   Published Jul 30, 2026self-described: Client engineering team, several weeks across dozens of pull requests

4. Mixed

   esposem, Agent Sandbox contributor

   [Discussion] Feedback for AI pre-reviewer (Coderabbit / Copilot) (kubernetes-sigs/agent-sandbox issue 1231)GitHub issue comment

   Finds CodeRabbit useful and mostly correct, but it drifts outside a pull request's scope, drops a finding as soon as a person disagrees, and can leave contributors talking only to a bot.

   Published Jul 21, 2026self-described: Kubernetes Agent Sandbox and other repositories that run CodeRabbit

5. Negative

   negz (Nic Cope), Crossplane maintainer

   Switch from CodeRabbit to copilot reviewsGitHub issue

   Set up CodeRabbit on Crossplane because it remembered feedback from earlier reviews; it works fine but is full of self-promotion and emojis, so he proposed GitHub Copilot review. Maintainers trialed Copilot on every pull request and then hit CodeRabbit rate limits.

   Published Jul 13, 2026self-described: Crossplane, a CNCF project; CodeRabbit chosen for remembering feedback from earlier reviews

6. Negative

   sensanaty, Engineer at a large public company

   Comment on: Claude Sonnet 5Hacker News comment

   At their ~2000-developer company, some teams, including theirs, are begging to disable CodeRabbit and similar MR bots because walls of noise make reviewing a nightmare.

   Published Jul 1, 2026self-described: Mature public company, ~2000 developers, large monorepo, broad AI tooling budget

7. Negative

   eptcyka, Hacker News commenter

   Comment on: When someone says they hate your productHacker News comment

   Their team tried CodeRabbit, but it enabled beta features unasked, such as AI replies on GitHub issues, and they disabled it; some teammates felt forced to use it.

   Published Dec 29, 2025self-described: Team

8. Mixed

   Mark Di Marco, startup hacker from Austin

   Picking an AI Code ReviewerPersonal blog post, markmarkoh.com

   A five-person team trialed CodeRabbit, Cubic, Cursor's BugBot and Claude Code Review side by side. CodeRabbit caught real bugs, but its nitpicks and refactor suggestions meant the team would ignore 75% or more of its comments, so they chose Cubic.

   "While it caught real issues, it also has a lot of “Nitpicks” and “Refactor Suggestions”, so we’d ignore 75% or more of the comments it made."

   Published Oct 10, 2025self-described: Five-person startup engineering team trialing four AI reviewers with minimal customization

9. Mixed

   Ahmed Atwa (nightknighto), AsyncAPI Generator contributor

   change coderabbit's config to be less annoyingGitHub issue

   Called CodeRabbit generally useful but too verbose and repetitive for small PRs; maintainers agreed to keep the summary and turn off the poem, suggested labels and reviewers, and make diagrams optional.

   Published Apr 24, 2025self-described: AsyncAPI Generator open source project, small frequent PRs

10. Mixed

    Yehonal, AzerothCore maintainer

    Proposal: coordinated bot reviews and PR risk triageGitHub issue

    Found CodeRabbit well configured for AzerothCore, but its reviews never become an approval, so maintainers still read every bot comment themselves. Another maintainer replied that the project is already overwhelmed by bot output.

    Published Aug 7, 2026self-described: AzerothCore open source project with 133 open pull requests, reviewed by CodeRabbit and Copilot

11. Negative

    anee769, Hacker News commenter

    Comment on: Show HN: Ox, an AI agent that catches tech debt before it's committedHacker News comment

    Their company ran CodeRabbit on Azure DevOps. It flagged many issues, most of them false positives, so the team stopped relying on it.

    Published Jul 1, 2026self-described: Company running CodeRabbit on Azure DevOps

12. Mixed

    stbenjam, Red Hat engineer

    CodeRabbit: Recommended setGitHub issue

    Asked Red Hat Product Security for a smaller must-have set of CodeRabbit checks. The product caps custom checks at 20, several duplicate what CodeRabbit or other scanners already do, and the AI-attribution check flags nearly every Claude pull request.

    Published May 26, 2026self-described: Red Hat Product Security's shared CodeRabbit configuration for other teams

13. Mixed

    rparolin, NVIDIA cuda-python maintainer

    Enable CodeRabbit auto-review on cuda-python (port CCCL's .coderabbit.yaml)GitHub issue

    The cuda-python team had turned automatic review off because the default profile was noisy. It proposed turning it back on with the tuned config NVIDIA's CCCL team adopted to suppress walkthroughs, poems and finishing-touch suggestions.

    Published May 25, 2026self-described: NVIDIA cuda-python repository, CodeRabbit installed in opt-in mode

14. Mixed

    friday, Ulauncher maintainer

    Tech Debt: Codebase Quality Improvements - CodeRabbit Baseline ScanGitHub issue comment

    Said CodeRabbit's baseline scan of Ulauncher listed mostly superficial issues the maintainers already knew about and had suppressed on purpose. They asked it for the logic bugs they had hoped it would find.

    Published Dec 19, 2025self-described: Ulauncher open source project, baseline scan of the whole codebase

15. Positive

    Maciej Korolik, Senior Frontend Developer and AI Expert at Monterail

    Best AI Code Review Tools 2026: Comparison & GuideCompany engineering blog post, Monterail blog

    Monterail ran a hands-on evaluation of GitHub Copilot, Cursor BugBot and CodeRabbit on an internal project and picked CodeRabbit, mainly for its PR summaries and sequence diagrams. Out of the box it posted many comments, and Monterail budgets one to two weeks of .coderabbit.yaml tuning.

    "At Monterail, we ran a formal hands-on evaluation of three leading tools: GitHub Copilot, Cursor BugBot, and CodeRabbit. CodeRabbit won."

    Published Oct 30, 2025Monterail, a software development company; GitHub Copilot, Cursor BugBot and CodeRabbit tested on one active internal project

16. Negative

    Orion Cohen (orionarcher), TorchSim maintainer

    Turn off Coderabbit?GitHub issue

    Found CodeRabbit really annoying and not particularly valuable and proposed turning off automatic reviews; the other maintainers agreed to keep it installed for on-demand reviews only.

    Published Sep 16, 2025self-described: TorchSim open source project, small maintainer group

17. Negative

    tnolet, Hacker News commenter

    Comment on: How we exploited CodeRabbit: From simple PR to RCE and write access on 1M reposHacker News comment

    Their team removed CodeRabbit because it was too verbose, produced too many false positives and never added anything useful.

    Published Aug 19, 2025self-described: Team (reply to a comment praising CodeRabbit)

18. Negative

    elsirion, Fedimint maintainer who enabled CodeRabbit

    Comment on: CodeRabbit AI PR reviewsGitHub issue comment

    Enabled CodeRabbit on Fedimint in 2024 with only the poem turned off, then removed it in April 2025 because everyone on the project got annoyed by it.

    Published Apr 20, 2025self-described: Fedimint open source project

Discussed without concern, 5

1. Positive

   Janet Kuo (janetkuo), SIG Apps chair and Agent Sandbox maintainer at Kubernetes

   REQUEST: New AI review tool integration for agent-sandbox (kubernetes/org issue 6296), 90-day evaluationGitHub issue comment, 90-day evaluation under the Kubernetes AI code review tools policy

   Agent Sandbox's 90-day review found 92% of CodeRabbit's 1,117 threads resolved or fixed and only 6% pushed back, while 15% of bot-approved PRs still drew human change requests. All ten contributors asked want to keep it, and she recommends continuing while capping comment volume on large PRs and tuning AGENTS.md.

   "Ten people gave an opinion; all ten want to keep it."

   "Continue on agent-sandbox. First-pass defects, missing test updates and API-convention slips get caught before a human looks…"

   Published Oct 15, 2026Kubernetes Agent Sandbox subproject, 740 PRs from 100+ authors in 90 days, CodeRabbit as first-pass pre-reviewer

2. Mixed

   Federico Capoano (nemesifier), OpenWISP maintainer

   [ci] Opt in valid PRs to continuous CodeRabbit reviewsGitHub issue

   Because CodeRabbit reviewed every PR, including ones later flagged invalid, the maintainer turned off automatic reviews and had OpenWISP's bot request reviews only on valid PRs.

   Published Aug 31, 2026self-described: OpenWISP open source repositories

3. Positive

   Janet Kuo (janetkuo), Agent Sandbox maintainer

   [Discussion] Feedback for AI pre-reviewer (Coderabbit / Copilot) (kubernetes-sigs/agent-sandbox issue 1231)GitHub issue comment

   Prefers CodeRabbit's review experience to Copilot's: it edits its existing comments without reposting them, is configurable, and learns when corrected, though Copilot catches some issues CodeRabbit misses.

   Published Jul 20, 2026self-described: Kubernetes Agent Sandbox project, running CodeRabbit and Copilot as pre-reviewers

4. Positive

   Jörg Thalheim (Mic92), Nix team member

   Coderabbit github app for the Nix repoGitHub issue

   Requested CodeRabbit for the Nix repository as an opt-in reviewer after it found bugs in other projects the requester maintains. It later caught use-after-free bugs before they merged, while others objected to its poems and comment volume.

   Published Oct 27, 2025self-described: NixOS Nix repository, opt-in reviews per pull request

5. Positive

   Benoît Viguier (ildyria), Lychee maintainer

   AI-assisted reviews, one month later...Project blog post, Lychee blog

   Categorizing all 290 CodeRabbit findings on 28 PRs over a month, the maintainer rated 72% relevant and 28% noise or wrong assumptions, with real catches including Zip Slip and IDOR flaws.

   Published Sep 13, 2025self-described: LycheeOrg open source projects on CodeRabbit's open source plan

GILD's fact checks, which don't count as sources

Checked Oct 12, 2026.CodeRabbit defaults to chill. The quiet profile puts only the most important comments inline and groups the rest in the summary. The assertive profile adds feedback the docs describe as nitpicky. The poem is off and the walkthrough collapsed by default. Every push starts an incremental review unless you turn off auto_incremental_review. Configuration reference

Counted for Engineering leaders and teams.

Raised as a concern, 4

1. Negative

   anee769, Hacker News commenter

   Comment on: Show HN: Ox, an AI agent that catches tech debt before it's committedHacker News comment

   Their company ran CodeRabbit on Azure DevOps. It flagged many issues, most of them false positives, so the team stopped relying on it.

   Published Jul 1, 2026self-described: Company running CodeRabbit on Azure DevOps

2. Positive

   Facundo F, CTO at DeployHQ

   AI Code Review Before You Deploy: Our Experience with CodeRabbitEngineering blog post, DeployHQ blog

   After more than a year as a customer, DeployHQ finds CodeRabbit's analysis sharper and more context-aware, though its security checks are sometimes overzealous, and pairs it with Claude Code and Codex.

   Published Dec 3, 2025self-described: DeployHQ, a CodeRabbit customer for over a year

3. Negative

   tnolet, Hacker News commenter

   Comment on: How we exploited CodeRabbit: From simple PR to RCE and write access on 1M reposHacker News comment

   Their team removed CodeRabbit because it was too verbose, produced too many false positives and never added anything useful.

   Published Aug 19, 2025self-described: Team (reply to a comment praising CodeRabbit)

4. Negative

   Martin Zoeller, Independent consultant who runs workshops on introducing AI into development teams

   CodeRabbit AI Review: Is It Worth It for Your Team?Independent review, MartinZoeller.com

   After weeks at a client, found reviews fast and some real frontend bugs caught. Noise, invented rules, wrong learnings and shallow context still made it not worth it next to a custom review skill.

   Published Jul 30, 2026self-described: Client engineering team, several weeks across dozens of pull requests

Discussed without concern, 2

1. Positive

   The_Fox, Hacker News commenter

   Comment on: There is an AI code review bubbleHacker News comment

   Installed CodeRabbit for their GitLab reviews; at roughly US$15 per user per month it regularly finds subtle problems, though it flags theoretical issues they just resolve.

   Published Jan 26, 2026self-described: Team using GitLab

2. Positive

   Benoît Viguier (ildyria), Lychee maintainer

   AI-assisted reviews, one month later...Project blog post, Lychee blog

   Categorizing all 290 CodeRabbit findings on 28 PRs over a month, the maintainer rated 72% relevant and 28% noise or wrong assumptions, with real catches including Zip Slip and IDOR flaws.

   Published Sep 13, 2025self-described: LycheeOrg open source projects on CodeRabbit's open source plan

Counted for Engineering leaders and teams.

Raised as a concern, 4

1. Mixed

   JonghunYu, Allra Fintech engineer

   Request CodeRabbit reviews for ready PRs and future ready changesGitHub issue

   Keeps CodeRabbit as a required review gate; several review requests came back as rate-limit refusals, and after retries its findings were fixed in follow-up pull requests before merge.

   Published Oct 21, 2026self-described: Allra Fintech texio repository, CodeRabbit as a required review gate

2. Negative

   urisland, Island engineer

   CodeRabbit still auto-reviews every PR despite the switch to Cursor BugbotGitHub issue

   After the repository replaced CodeRabbit with Cursor Bugbot, the org-wide app kept reviewing every pull request on default settings. It posted rate-limit notices of 2 reviews per hour until the maintainer uninstalled it.

   Published Sep 6, 2026self-described: island-io/mila repository, CodeRabbit installed on the island-io GitHub organization

3. Negative

   lukeasrodgers, Hacker News commenter

   Comment on: Open Code Review - An AI-powered code review CLI toolHacker News comment

   Their team tried CodeRabbit and Qodo and found both far worse than a thin in-house Claude/Codex wrapper, arguing most teams should not pay SaaS fees for AI review.

   Published Jun 5, 2026self-described: Team

4. Negative

   negz (Nic Cope), Crossplane maintainer

   Switch from CodeRabbit to copilot reviewsGitHub issue

   Set up CodeRabbit on Crossplane because it remembered feedback from earlier reviews; it works fine but is full of self-promotion and emojis, so he proposed GitHub Copilot review. Maintainers trialed Copilot on every pull request and then hit CodeRabbit rate limits.

   Published Jul 13, 2026self-described: Crossplane, a CNCF project; CodeRabbit chosen for remembering feedback from earlier reviews

Discussed without concern, 4

1. Positive

   TheAnkurTyagi, Hacker News commenter

   Comment on: Ask HN: How much are you guys paying for AI coding tools monthly?Hacker News comment

   Lists CodeRabbit at US$30 a month among four paid AI tools, under a team soft cap of US$500 per developer per month that most developers stay well below.

   Published Sep 1, 2025self-described: Team with a US$500 monthly AI tooling cap per developer

2. Negative

   Martin Zoeller, Independent consultant who runs workshops on introducing AI into development teams

   CodeRabbit AI Review: Is It Worth It for Your Team?Independent review, MartinZoeller.com

   After weeks at a client, found reviews fast and some real frontend bugs caught. Noise, invented rules, wrong learnings and shallow context still made it not worth it next to a custom review skill.

   Published Jul 30, 2026self-described: Client engineering team, several weeks across dozens of pull requests

3. Positive

   The_Fox, Hacker News commenter

   Comment on: There is an AI code review bubbleHacker News comment

   Installed CodeRabbit for their GitLab reviews; at roughly US$15 per user per month it regularly finds subtle problems, though it flags theoretical issues they just resolve.

   Published Jan 26, 2026self-described: Team using GitLab

4. Positive

   Benoît Viguier (ildyria), Lychee maintainer

   AI-assisted reviews, one month later...Project blog post, Lychee blog

   Categorizing all 290 CodeRabbit findings on 28 PRs over a month, the maintainer rated 72% relevant and 28% noise or wrong assumptions, with real catches including Zip Slip and IDOR flaws.

   Published Sep 13, 2025self-described: LycheeOrg open source projects on CodeRabbit's open source plan

GILD's fact checks, which don't count as sources

Checked Oct 12, 2026.Essentials includes 5 PR reviews per developer per hour. Team includes 8, Advanced 10, and Enterprise 12. Open source projects get 1 to 10, depending on the project. At the limit, admins can pause reviews or pay US$0.25 per reviewed file under a monthly cap. Plans and pricing