Yes. GitHub builds one in: Copilot code review, which you request from a pull request's Reviewers menu or set up to run automatically. Third-party reviewers work through their own GitHub App on the repositories you choose (Claude Code Review, CodeRabbit, Cursor Bugbot, Graphite, Greptile, Qodo), Codex reviews repositories you connect to Codex, and the Claude Code GitHub Action runs inside your own workflow. For PR review on GitHub, these AI code review tools differ in what starts a review and what they leave on the pull request. Most of them take a command in a pull request comment, such as @coderabbitai review or @codex review. Claude Code Review, CodeRabbit, Cursor Bugbot and Greptile also document a check or status that their review posts on the pull request.
Each fact here is from the vendor's own pages, read on 5 November 2026; no vendor saw this text before publication.
AI reviewers that comment on GitHub pull requests
This list covers the reviewers whose GitHub setup pages I read on 5 November 2026, each cell in the vendor's words or a close paraphrase; a tool missing here is not a judgment about it.
| Reviewer | How it gets onto GitHub | What starts a review | What it leaves on the pull request |
|---|---|---|---|
| GitHub Copilot code review | part of GitHub: on a pull request, under Reviewers in the right sidebar, click Request next to Copilot | "By default, Copilot only reviews a pull request if you assign it to the pull request." and "However, you can configure automatic reviews." | review comments, each labeled High, Medium or Low severity, where possible with "suggested changes which you can apply with a couple of clicks"; a Comment review by default |
| Claude Code Review (Anthropic) | the Claude GitHub App, on the repositories an Owner adds in Claude's admin settings; a separate path, the Claude Code GitHub Action, "runs Claude Code inside your repository's workflows" | a Review runs setting for each repository: "Each repository starts set to review after every push, which runs the most reviews."; in Manual mode "opening or pushing to a pull request doesn't start a review."; in any mode, an @claude review comment from someone with write access asks for one | inline comments, a summary in the review body and "a check run named" Claude Code Review; "A review never approves or blocks your pull request, and the check run never concludes as a failure." |
| CodeRabbit | installed from CodeRabbit after Login with GitHub, on the organization you pick, for all repositories or only selected ones | by default, pull requests to the default branch, and "By default, CodeRabbit skips draft PRs."; whatever the automatic settings, an @coderabbitai review comment starts one | a pull request summary and review comments; on GitHub with CodeRabbit Cloud, a check run named CodeRabbit while reviews.review_progress is true, its default |
| Codex (OpenAI) | a GitHub repository connected to Codex, with code review turned on for it in Codex settings | @codex review in a pull request comment, or automatic reviews once Automatic review is turned on in Codex settings, where you "Choose when reviews run with" Review trigger | "Codex posts a review on the pull request, just like a teammate would."; in GitHub, "Codex flags only P0 and P1 issues" |
| Cursor Bugbot | the Cursor GitHub app, connected from the Cursor dashboard | "It runs automatically on each PR update or manually when triggered.", manually by a cursor review or bugbot run comment | comments "with explanations and fix suggestions", a summary added to the pull request description by default, and a status for each run, which on GitHub "appears as a check named" Cursor Bugbot |
| Graphite | repositories selected on Graphite's AI code review settings page; for personal repositories, the Graphite GitHub App on your own account | automatically, "all new pull requests in the selected repositories" | Graphite Agent "Comments directly on pull requests when it finds potential issues" and suggests fixes |
| Greptile | the "Greptile Apps GitHub App" for github.com, with repositories enabled in Greptile | the autoReview setting ("Events that start a review unasked:"), open by default, which covers opened, reopened, ready for review and a trigger label; on request, you "trigger reviews with @greptileai" | "PR summaries, confidence scores, inline comments, suggested fixes, and diagrams"; a review status check, with statusCheck on by default |
| Qodo | Qodo on GitHub, from the docs page "Install Qodo on GitHub" for "GitHub Cloud, multi-tenant." | "Automatic is the default. Qodo runs a review as soon as a pull request is opened, reopened, or marked ready for review, without any manual trigger."; on request, a /agentic_review comment | "Feedback appears based on your configuration (summary, inline, or both)" |
| Sigma (Mnemoverse, the author's company) | "Create an account. No invitation needed." (https://sigma.mnemoverse.com, read on 5 November 2026); "Connect GitHub once; reuse your saved connections." | "Prefer GitHub? Comment /sigma review on the PR."; its docs: "On request starts reviews only when asked; Automatic can also start them after PR events." | its docs: "a neutral check per reviewer on the head commit, which tracks the run, and a review with the findings" |
How a reviewer gets onto GitHub
The first eight reviewers in the table reach GitHub in four ways: built into GitHub, through the vendor's own GitHub App on the repositories you choose, through repositories connected to Codex, or as an Action inside your own workflow.
Copilot code review is part of GitHub. On a pull request, under Reviewers in the right sidebar, you click Request next to Copilot, and GitHub notes that "Copilot code review uses GitHub Actions to run agentic capabilities."
The third-party reviewers each name their own starting point. At Anthropic, an Owner of the Claude organization turns on Code Review in its admin settings to "connect GitHub, install the Claude GitHub App, choose repositories, and set when reviews run", and "Reviews run on Anthropic infrastructure", with no workflow files added to the repositories. CodeRabbit starts with Login with GitHub; the next step reads "Choose the organization where you want to install CodeRabbit.", followed by access to all repositories or only selected ones. Cursor starts in its dashboard: "Connect your repositories through the Cursor dashboard to start using Bugbot." The app behind it is Cursor's own ("The Cursor GitHub app connects your repositories so you can use features like" Cloud Agents and Bugbot), and "On GitHub, connecting turns Bugbot on for the repositories you grant access to." Graphite turns reviews on from its AI code review settings page, where you "Select the repositories where you want AI reviews enabled", and for personal repositories "you must first install the Graphite GitHub App on your personal account." Greptile's steps are to "install the GitHub App or add a GitLab token, link your organization or group, and enable repositories." Qodo's docs index links to "Install Qodo on GitHub", and its review page says "Qodo works directly inside your Git provider and integrates into your existing pull request workflow."
Codex works from repositories connected to Codex: "Connect your GitHub repository to Codex.", then in Codex settings "Choose your repository, then turn on" Automatic review under Review code. That step needs a connected repository and "GitHub push or admin permission for its settings."; each developer can also turn on Automatic review for their own pull requests under Personal preferences.
The Claude Code GitHub Action is a separate path from Anthropic's managed Code Review: it "is a GitHub Action that runs Claude Code inside your repository's workflows." Its quick setup runs /install-github-app from Claude Code, and then "Claude Code installs the GitHub App, adds your authentication secret, and prepares the workflow pull request for you". With the review workflow selected, "Claude posts each review on the pull request itself, as an inline comment on each issue it finds or as one summary comment when it finds none."
Who sets the reviewer up decides the path: whoever administers the GitHub organization or the repository connects an App or a repository, and a team that wants the reviewer inside its own CI looks at the Action.
What starts a review
Some reviewers take every new pull request by default, and others wait for a request or for a switch in their settings. Most of them also accept a command in a pull request comment, and the table above gives each one. On a busy repository, starting with a manual mode or that comment command, where the vendor offers one, limits reviews to the pull requests someone asks about. Each vendor's default stands next to its prices in Alternatives to Copilot for Code Reviews?.
What stays on the pull request
Claude Code Review, CodeRabbit, Cursor Bugbot and Greptile also document a check or status that their review posts on the pull request. Each of the first eight reviewers in the table leaves its findings on the pull request itself, as comments, a GitHub review or both. Copilot leaves a Comment review by default, and "Copilot's review comments behave like review comments from humans." Codex posts a standard GitHub review: "Codex posts a review on the pull request, just like a teammate would." Graphite Agent "Comments directly on pull requests when it finds potential issues", and at Qodo "Feedback appears based on your configuration (summary, inline, or both)".
Each of the four names its check in its own words. Claude Code Review posts "a check run named" Claude Code Review. "CodeRabbit reports each review on the pull request's head commit under the name" CodeRabbit, as a check run on GitHub with CodeRabbit Cloud while reviews.review_progress is true, its default. Bugbot's status "appears as a check named" Cursor Bugbot. Greptile's statusCheck setting, true by default, is there to "Post a review status check on GitHub, GitLab, Bitbucket (Cloud and Data Center), Gitea, and Origin."
On branch rules, in each vendor's words: Anthropic writes "It never concludes as a failure, so a finished review doesn't fail a required check." CodeRabbit writes "Branch rules can require it before a pull request merges." Cursor writes "If you use branch protection, require the Bugbot check or build status to make sure Bugbot runs before merge.", adds "Requiring the status alone does not block merges on findings because findings default to" neutral, and then "If fail-on-unresolved-issues behavior is available for your organization, enable it to make unresolved findings produce a failing status."
Which setup fits which team
Take a team that already writes code with an AI editor and wants a second pass on its GitHub pull requests before merge. If it already pays for Copilot, the first place to look is the Reviewers menu of a pull request, where its plan and policies allow Copilot code review. If the review should count toward merging, look at the reviewers that document a check or status, and at what each vendor says a branch rule does with it. If the reviewer should live in the team's own CI, the Claude Code GitHub Action runs inside its workflows. Prices and trial options are compared in Alternatives to Copilot for Code Reviews?, and how to count which reviewer is right on your own pull requests is in a comparison of what AI code reviewers say they check before commenting.
Sources and method
All pages listed here were read on 5 November 2026; every quotation matched its page word for word on that day, and the text describes those pages as they stood then. The list is not exhaustive: it covers the vendors whose pages are linked below.
GitHub: Copilot code review, using Copilot code review
Anthropic: setting up Code Review, Code Review on pull requests, Claude Code GitHub Actions
CodeRabbit: GitHub, automatic reviews, status check
OpenAI: Codex code review in GitHub
Cursor: Bugbot, GitHub integration
Graphite: AI reviews setup
Greptile: code providers, configuration reference, docs index
Qodo: triggering a review, docs index
Sigma: docs
