36Code reviewProvisional score

Sourcery

AI code reviewer built for the agent era that catches bugs and enforces standards in pull requests.

Updated July 31, 2026

The read

Sourcery reviews pull requests with a focus on code written by AI coding agents, detecting bugs, security vulnerabilities, and style violations and returning line-by-line feedback inside GitHub and GitLab. It also provides real-time feedback in VS Code and JetBrains during writing, so issues surface before a PR is opened. Three hundred thousand developers use the platform, which is SOC 2 certified and offers a zero-code-retention option. A free tier for public repositories is available, and a Pro plan for private repositories costs 12 dollars per seat per month. The free tier is limited to three public repositories scanned biweekly, restricting usefulness for open-source teams with many repos.

Where it shines

Zero-code-retention option available for teams with strict data privacy requirements
Real-time IDE feedback catches issues before a PR is even opened
SOC 2 certified with 300k-plus developer user base

Worth knowing first

Free tier covers only three repos scanned biweekly, limiting open-source teams
GitLab and GitHub only, no Bitbucket or Azure DevOps integration mentioned

Index score

74.2

Provisional. This score is read from the product's public capability, not from the one-prompt rebuild the fully reviewed entries went through, so treat it as a placing rather than a verdict.

Craft74

Speed75

Control72

Value76

Rank: 36 of 54
Builds: Tools
Category: Code review
Output: Inline PR review comments with IDE real-time feedback
Pricing: Free tier
Best for: Teams wanting AI review specialized for AI-generated code quality

Scores are our own editorial judgement, weighted craft 30, speed 25, control 25 and value 20.

Also worth a look

Full index