Starter
Try AI PR review.
- 2 reviewers
- Code Review
- Security Review
- 3 seats included
- 15,000 credits to start, then 5,000/mo
- About 70 reviews to start, then about 25 a month
- Unlimited repos
Cursor Bugbot reviews pull requests and comments with likely bugs and fixes. Since May 2026 it is billed by usage, with no seat fee: Cursor puts the average review at $1.00–$1.50, paid from included usage on Premium, Super and Ultra, then prepaid credit or on-demand usage, and from prepaid credit on Teams.[1][2] The Free and Plus plans do not include it. What a team pays for Bugbot.
CriticalKnown issue in the Code Review benchmark · caught by DeepSeek V4.1 FlashSee the graded case
On the same 50 open-source pull requests, Bugbot found 74 of 158 known issues and IonWarp's Code Review 77.[3] Bugbot had fewer findings that matched none, 56 to 78, but missed a race in Cal.com's new backup codes.[4] IonWarp's Starter plan starts with about 70 reviews, then Pro is $49/mo for 5 seats. Bugbot vs CodeRabbit runs the same test.
# .cursor/BUGBOT.md
- Flag SQL built from strings.
- New API routes check the session first.
- Skip generated files in src/gen/.Connect your repositories in the Cursor dashboard and pick the ones Bugbot reviews under Automations. Outside GitHub, you turn it on per repository.[5]
It reviews every pull request update, by default only what changed since its last review. Comment cursor review or bugbot run to start one.[5]
It reads .cursor/BUGBOT.md at the root, and one in a folder for changes under it; .cursor/rules files do not apply. Comment @cursor remember to save a learned rule.[5]
Fix in Cursor opens the fix in the editor; Autofix pushes it from a Cloud Agent, billed as Cloud Agent usage.[5]
| Criterion | ||
|---|---|---|
| Cost | $49/mo for 5 developers — one Pro workspace with 5 seats included. | $100–$150/mo for 5 developers at 100 reviews a month — $1.00–$1.50 a run.[1] |
| Git platforms | GitHub, and GitHub Enterprise on the Enterprise plan. | GitHub, GitLab, Bitbucket and Azure DevOps.[5] |
| Separate reviewers | Yes:16 reviewers, such as Security Review, SOC 2 Review and Docs Review. The planner adds the ones each diff needs. | One review covers bugs, security issues and code quality; a separate Security Reviewer agent checks pull requests.[5][6] |
Competitor facts are from each vendor's own pricing page and docs, dated in the sources below. IonWarp's column is read from its own pricing and reviewer catalog.
Real IonWarp comments on open-source pull requests, graded against the benchmark's answer key.
Missed byCursor Bugbot 21
3 of 21 · See the benchmark
getSubGroupsCount returns null when cached group is deleted concurrently, violating the Never-null contract of GroupModel · Code Review
CriticalKnown issue in the Code Review benchmark · caught by Muse Spark 1.3See the graded case
Public members template iterates bare members while guarding model - template guards on model but loops bare members, and route sets controller model only, so list can render empty · Code Review
HighKnown issue in the Code Review benchmark · caught by Muse Spark 1.3See the graded case
Legacy DeleteCollection outcome goes to recordStorageDuration instead of recordLegacyDuration, hiding legacy failures from legacy SLOs · Code Review
HighKnown issue in the Code Review benchmark · caught by Muse Spark 1.3See the graded case
Muse Spark 1.3 is its best model on this benchmark; Code Review runs DeepSeek V4.1 Flash unless you pick another.
Missed byCursor Bugbot 6
2 of 6 · See the benchmark
P2Single-use guarantee of backup codes is not concurrency-safe (TOCTOU) · Security Review
When two login requests supply the same backup code concurrently, both read the stored array before either write lands, so both pass indexOf and both log in. This breaks the UI promise in the new locale string backup_code_instructions: 'Each backup code can be used exactly once'. Trigger: parallel signIn calls with the same code (e.g. double-clicked submit).
HighKnown issue in the Security Review benchmark · caught by GLM 5.3 FlashSee the graded case
P2Denial cache key concatenation is ambiguous and can collide across (name, parent) pairs · Security Review
When a user's deny result is cached, userPermDenialCacheKey joins name and parent with '_' without escaping, so Check(name="a", parent="b_c") and Check(name="a_b", parent="c") produce the same key; a deny computed for one object denies the other for up to shortCacheTTL (30s). Over-denial only — no privilege gain — and requires two names differing only by underscore placement, so cosmetic-to-minor.
MediumKnown issue in the Security Review benchmark · caught by GLM 5.3 FlashSee the graded case
GLM 5.3 Flash is its best model on this benchmark; Security Review runs DeepSeek V4.1 Flash unless you pick another.
Missed byCursor Bugbot 4
1 of 4 · See the benchmark
P2Generic Error thrown on missing credential misstates client error as 500 server error · Log & Errors Review
When credential is not found or not owned by user, throwing a generic Error causes tRPC to respond with 500 INTERNAL_SERVER_ERROR instead of a 404 NOT_FOUND error contract.
HighKnown issue in the Log & Errors Review benchmark · caught by Gemini 3.8 FlashSee the graded case
Gemini 3.8 Flash is its best model on this benchmark; Log & Errors Review runs DeepSeek V4.1 Flash unless you pick another.
Missed byCursor Bugbot 5
1 of 5 · See the benchmark
P2Log message refers to login on 2FA disable endpoint · Docs Review
The error log states 'cannot proceed with backup code login', but this handler is the 2FA disable endpoint (/api/auth/two-factor/totp/disable), misleading operators diagnosing disable failures.
LowKnown issue in the Docs Review benchmark · caught by Gemini 3.8 FlashSee the graded case
Gemini 3.8 Flash is its best model on this benchmark; Docs Review runs DeepSeek V4.1 Flash unless you pick another.
One workspace price with seats included. Credits pool across your team. Unlimited repos on every plan.
Try AI PR review.
Get AI PR review right from day 1.
Scale AI PR review.
Custom reviewers + credits · GitHub Enterprise · SSO/SAML + audit logs · Custom security review + SLA
Cursor bills Bugbot by usage, with no seat fee, and puts the average review at $1.00–$1.50, depending on pull request size and complexity (Cursor, Nov 2026). On Premium, Super and Ultra, reviews come out of included usage first, then prepaid credit or on-demand usage. On Teams they bill from prepaid credit.
No. Cursor's Free and Plus plans do not include Bugbot, and its Bugbot page lists a trial only for Enterprise teams, through their account team.
Turn it on for the repository under Automations in the Cursor dashboard, and it reviews each pull request update. To start a review by hand, comment cursor review or bugbot run on the pull request.
Bugbot is one review billed per run. IonWarp runs Code Review on every pull request, adds reviewers for security, performance, docs and architecture when the diff needs them, and can run next to Bugbot on the same pull request.
By F1 on the same 50 pull requests and 158 known issues: Qodo 58.5%, Greptile 51.1%, IonWarp's Code Review 49.2% and CodeRabbit 42.2%, against Cursor Bugbot's 51.4%. Graphite, Cubic and Anthropic's Claude Code Review are other AI code reviewers, not in our benchmark.
getSubGroupsCount() returns null when the cached group was deleted in another session and the lazy delegate resolves to null, contradicting GroupModel's documented never-null contract and causing admin REST responses to receive null counts · Code Review