CodeRabbit

Hosted AI code reviewer for pull requests on GitHub, GitLab, Azure DevOps and Bitbucket, with CLI and IDE reviews before you push.

Freemium#ai#code-review#pull-requests#security

Visit website ↗Documentation

What is CodeRabbit?

CodeRabbit is a hosted AI code reviewer. Installed on a Git organization, it reviews each pull request when it opens and, by default, again on every new push, posting a summary, a walkthrough and line comments you can reply to or ask it to fix. It works on GitHub, GitLab, Azure DevOps and Bitbucket, cloud or self-managed. The same reviewer runs locally through a CLI (0.9.5 on 11 November 2026) and extensions for VS Code-based editors and JetBrains IDEs (beta).

CodeRabbit Inc. is closed source and venture-backed; it raised a $143M Series C at a $1.5B valuation on 12 August 2026. It charges per developer who opens pull requests, with hourly review allowances, and public repositories are free. Self-hosting is Enterprise-only and needs 500+ seats.

Since July 2026 it has added security scanning (CodeRabbit Security), a cross-repo PR queue (Triage) and an Advanced plan, and renamed Pro and Pro+ to Essentials and Team. On 12 November 2026 it said it had passed $100M in annual recurring revenue. Its main rivals are Greptile and the reviewers built into GitHub Copilot and Claude Code.

Our review

3.5

The safe default if you want a dedicated PR reviewer: it runs on every major Git host, the per-seat bill is predictable, public repositories are free, and the same reviewer runs in your editor and terminal before you push. The catch is review quality. Developers in the last 90 days complain more about noise and shallow comments than they praise its findings, and we haven't run it on our own repositories. Heavy, agent-driven teams should also read the fair-usage rules: allowances slow down after about 30 reviews per developer per week, and overage is billed per file.

Pros

+ Widest host support in the category: GitHub, GitLab, Azure DevOps, Bitbucket Cloud and Data Center, including self-managed instances
+ Predictable seat pricing ($24–$72 per developer per month, annual) that only counts developers who open PRs; public repositories are free with Team features
+ Reviews before the PR too: the CLI and IDE extensions run the same reviewer locally, and the Free plan includes IDE and CLI reviews (3 per hour)

Cons

− The most common community complaint is noise: many low-value comments and few findings about design, architecture or domain logic
− Hourly allowances shrink under sustained use (fair-usage bands), and over-limit reviews cost $0.25 per reviewed file with no spending cap by default
− Closed source; self-hosting needs an Enterprise contract with 500+ seats

Reviewed by SeekDeploy Editorial · updated 2026-11-15. Reviews are independent — we don’t accept payment for coverage or ratings.

Should you use CodeRabbit?

A good fit if you're…

Teams on Azure DevOps or Bitbucket Data Center, or split across several Git hosts, who want one PR reviewer everywhere
Teams that want a fixed per-developer bill rather than paying per review
Open-source maintainers: public repositories get full reviews free
Developers who want the same reviewer in the editor or terminal (and in coding-agent loops) before they open a PR

Probably not for you if you're…

Teams that mostly want deep architectural or domain review; users report it's strongest on local, line-level issues
Organizations with many occasional contributors, where per-seat pricing wastes money
Companies that need self-hosting but have fewer than 500 seats

Pick something else if…

you'd rather pay per review than per seat (many occasional contributors), or want end-to-end test runs on PRs (TREX) → Greptile
you already pay for Copilot seats and only need PR review on GitHub → GitHub Copilot
you want to run reviews yourself with /code-review or in your own CI instead of buying another service → Claude Code

Gotchas to know before you commit

By default CodeRabbit reviews every push, not just the PR opening, and each push draws one review from the hourly allowance. Editing a file in GitHub's web UI or clicking Update branch counts too. Set auto_incremental_review to false to review only the opening push.
Fair usage: on Essentials the refill rate falls from 5 to 1 review per hour once a developer passes 60 reviews in 7 days (bands start at 30). Coding agents that push often hit this fast.
The trial switches usage-based reviews to Automatic, and the default monthly spending cap is Unlimited. Keep it on after the trial and over-limit reviews bill at $0.25 per reviewed file until an admin sets a cap.
With auto-approval, anyone who opens a PR gets a paid seat; ignore_usernames skips their reviews but doesn't stop the seat assignment.
In January 2025 Kudelski Security found RuboCop running outside CodeRabbit's sandbox, enough to run code from a PR on its servers. CodeRabbit says it fixed this within hours and no customer data was accessed (CEO post, 19 August 2025).

CodeRabbit pricing

Checked Nov 15, 2026 against the official pricing page

Per-seat subscription, charged only for developers who open pull requests: $24, $48 or $72 per developer per month billed annually ($30, $60 or $90 month to month). Each seat has an hourly review allowance that slows under sustained use; past it, reviews either pause or continue at $0.25 per reviewed file, so cost climbs with agent-driven PR volume and large PRs.

Free / Open source

Individuals, and anyone on public repositories

Private repos: PR summaries only, plus IDE and CLI reviews (3 per hour)
Public repos: full reviews with Team features

Limit: No PR reviews on private repositories
Limit: OSS rate limits vary by project; repos under 10 stars need a manual review trigger

Essentials

$24 / developer / month (annual), $30 monthly

Teams that want PR review

PR, IDE and CLI reviews
Learnings, one-click fixes, built-in pre-merge checks
Jira and Linear integration

Limit: 5 PR reviews per developer per hour, 150 files per review
Limit: Refill drops to 1 per hour after 60+ reviews per developer in 7 days
Limit: 1 linked repository for cross-repo context

Team

$48 / developer / month (annual), $60 monthly

Teams that want the workflow around review

Everything in Essentials
Triage PR queue, unit-test generation, merge-conflict resolution
Custom pre-merge checks

Limit: 8 PR reviews per developer per hour, 300 files per review

Advanced

$72 / developer / month (annual), $90 monthly

Teams that want security review on every PR

Everything in Team
Continuous PR security review, attack-surface map
Blast-radius and architecture-impact analysis

Limit: 10 PR reviews per developer per hour
Limit: Deep Scan of whole repositories billed separately

Enterprise

Custom

Large or regulated organizations

SSO, custom RBAC, audit logs, API
Self-hosting option, marketplace billing

Limit: Self-hosting requires 500+ seats

What it costs in practice

Solo developer	$0 or $24–30 / month Free covers IDE and CLI reviews (3 per hour) but only summaries on private PRs. PR reviews on a private repo need one Essentials seat: $24 annual or $30 monthly. Greptile's Starter gives 50 free PR reviews a month instead.
Team of 10, ~200 PRs/month	~$240–300 / month All 10 open PRs, so 10 Essentials seats: 10 × $24 = $240 (annual) or 10 × $30 = $300 (monthly). PR count doesn't change this: 20 PRs per developer a month is far below the 5-per-hour allowance. Team plan doubles it ($480–600). Compare Greptile's ~$200–350 for the same team.
Same team, reviews on every push	~$240–300 / month Every-push reviews are CodeRabbit's default, so this is the same bill. 3 reviewed pushes per PR = 600 review events, ~60 per developer a month or ~14 a week, under the first fair-usage band (30 a week on Essentials). Greptile's equivalent is ~$600 before TREX. Agent-heavy teams above ~30 reviews per developer a week get slower refills, or pay $0.25 per reviewed file if usage-based reviews are on (a 20-file review costs $5).

Estimates in USD from list prices. Your bill depends on usage, region, and discounts.

Health & activity

Updated Nov 15, 2026 · last 90 days

Closed sourceThe product itself isn't public. The figures below are for its public companion repo, so treat them as a partial view.

GitHub stars

7 forks

Issues opened / closed

8 / 3

Last push

Jan 30, 2026

Open issues & PRs

From coderabbitai/vscode-extension (Public issue tracker for the CodeRabbit VS Code extension (the review service itself is closed source)). Counts cover the last 90 days unless noted.

Downloads

npm · coderabbit4.4k last 30 days
Homebrew Cask · coderabbit1k installs, last 30 days

What the community says

Hacker News & GitHub issues · Aug 17, 2026 – Nov 15, 2026

Most of the conversation in this window is CodeRabbit's own announcements (a $10M+ open-source pledge, free security scans, new security agents, $100M ARR), plus passing mentions of CodeRabbit as the default name for a PR review bot. Developer comments are mixed: open-source maintainers and some teams rely on it, while engineers who have used it at work mostly complain about noise and shallow findings.

What people like

Free reviews and money for open source · 3 sources
Maintainers credit CodeRabbit for free reviews on public repositories and for cash sponsorship; the mise maintainer says he depends on it for code review, and Bun's PRs show it in their review loop.
HN comment 1 HN comment 2 HN comment 3
A useful extra automated check · 4 sources
Some developers keep it in their pipeline next to tests, CodeQL and SonarQube, and one prefers it over Sourcery, Greptile and Kodus. A consultant says client teams like it.
HN comment 1 HN comment 2 HN comment 3

What people complain about

Too much noise · 5 sources
Several engineers say the volume of low-value comments outweighs the useful ones: one team dropped it (and Qodo) for an internal tool, a former security researcher calls the friction not worth it, and others mention generated 'poetry' and an imperative tone.
HN comment 1 HN comment 2 HN comment 3
Shallow reviews that miss real problems · 5 sources
Users say it catches typos and local mistakes but not architecture, domain or design errors, and a local agent did better. One security write-up found vulnerabilities in code CodeRabbit had reviewed; another notes diff-anchored bots report a sample of a repeated bug, not every site.
HN comment 1 HN comment 2 HN comment 3
Easy to replace with your own agent · 4 sources
Several teams say a coding agent with a review skill, or an in-house bot triggered from webhooks, does as well for less money, and one objects to per-user pricing.
HN comment 1 HN comment 2 HN comment 3
Humans start trusting it too much · 2 sources
Commenters see developers treating its comments as gospel or reviewing less strictly once the bot has passed a PR.
HN comment 1 HN comment 2

Written by Idris Okonjo on Nov 15, 2026 after reading 34 relevant Hacker News comments, threads and GitHub issues from Aug 17, 2026 to Nov 15, 2026. Each point links to the discussions behind it, and the count shows how many sources raised it. This is a reading of public discussion, not a survey.