Bugbot
Cursor's AI code reviewer: automatic PR reviews with inline comments, severity levels, and one-click fixes.
About
Bugbot is Anysphere's automated pull-request reviewer, built as part of the Cursor ecosystem. It launched in beta alongside Cursor 1.0 in June 2025 and went generally available the following month. On every push to a pull request it reads the diff, pulls in surrounding repository context, and posts inline comments on lines that look risky, with severity levels and a 'Fix in Cursor' handoff that deep-links each issue into a Cursor agent session. It runs several analysis passes with majority voting and a validator model to cut false positives, per Anysphere's engineering write-up, and it is explicitly tuned to catch bugs in AI-generated code. Configuration lives in the Cursor dashboard, with natural-language Bugbot Rules for enforcing team coding standards.
Pricing moved to usage-based billing in mid-2026, roughly one to one-and-a-half dollars per run depending on PR size, on top of Cursor plan seats. Coverage has grown beyond the original GitHub-only scope to include GitLab and Bitbucket. The honest caveat is structural: Bugbot reads the diff, it does not run your code, so a finding is a confident prediction from the patch text, never a reproduced failure. That is fine for logic errors and common anti-patterns, and a real ceiling for runtime or authorization regressions. It is also SaaS-only with no self-hosted option and no bring-your-own-model.
What it can do
- Automatic PR review on every push with inline comments
- Severity levels and multi-pass analysis with false-positive filtering
- 'Fix in Cursor' handoff: deep-link findings into a Cursor agent session
- Natural-language Bugbot Rules for team coding standards
- GitHub, GitLab, and Bitbucket support; REST API for triggering reviews
Setup
- Install the Cursor GitHub App (or GitLab/Bitbucket integration) on your organization.
- Enable Bugbot for the repository in the Cursor dashboard Automations section.
- Reviews trigger automatically on PR open and push, or manually by commenting
bugbot run. - Review the inline comments and severity labels; use 'Fix in Cursor' to hand findings to a Cursor agent, and define Bugbot Rules in natural language for team standards.
Prerequisites
A Cursor plan (Individual, Pro, Pro Plus, Teams, Ultra, or Enterprise); admin access to enable Bugbot per repo; GitHub, GitLab, or Bitbucket.
Use cases
Works with
Trust score — 7/10
Scored 7/10 (verified tier, metrics-v1). Evidence: launch timeline (beta June 2025, GA July 2025) corroborated by multiple 2025-2026 sources; mid-2026 usage-based pricing (~$1-1.50/run) reported by competitor comparisons; GitLab/Bitbucket coverage and the /bugbot/review API documented in community engineering notes; cursor.com verified live. Strengths: deep Cursor ecosystem integration, active 2026 development, tuned for AI-generated code. Docked for: (1) never executes code — findings are predictions, not reproductions; (2) proprietary, SaaS-only, no BYO model. License: proprietary.
How the trust score is built
Full methodology →- Adoption · stars/downloads, log-scaled vs peers
- 7.0/10
- Maintenance · recency, releases, issue load
- 9.0/10
- Openness · license + docs
- 3.0/10
- Manual review · sanity read + red-flag check
- 7.0/10
- · Beta June 2025 with Cursor 1.0, GA July 2025 (multiple 2025-2026 sources)
- · Mid-2026 usage-based pricing ~$1-1.50/run reported by competitor comparisons
- · GitLab/Bitbucket coverage and /bugbot/review API documented in community engineering notes
- · cursor.com verified live; never executes code (predictions only); SaaS-only, no BYO model
Sources
Every fact above was checked against these sources. Where a claim comes from the provider and couldn't be independently verified, the text says so.