Revix AI

Revix AI is an AI code review product from founders Oscar Estudillo and Uri Gonfaus. It connects to GitHub, Bitbucket, or GitLab, reads the repository to extract review rules, and reviews each pull request against them. It also produces short pull request summaries with key changes and a risk level.

What It Is

Revix is an automated pull request reviewer. Instead of requiring a hand-written rulebook, it analyzes the codebase and writes rules from patterns already present, which teams can edit, add to, or disable. The FAQ distinguishes it from a linter: a linter checks syntax and formatting against fixed rules, while Revix checks whether a change follows the team's conventions, structure, and expectations.

How the review workflow works

Users connect a Git provider account, choose repositories, and Revix starts commenting on new pull requests. Comments follow a conventional-comments style, labeled by intent and severity so authors can prioritize fixes. Each pull request also gets a summary noting key changes and risk level. The site says it adapts to any language by reading code, conventions, and existing rules.

Privacy and organizational features

The FAQ states code is never used to train models, reviews run on isolated infrastructure, and source is purged after each run. The Enterprise tier is described as SOC 2 Type II and ISO 27001 certified, with roles, granular permissions, and Slack and Jira integrations available on higher plans.