Vercel Security Review is now in public beta for Pro and Enterprise teams. Once enabled, every push to a connected GitHub repository is reviewed for vulnerabilities, and the findings appear in the Vulnerabilities dashboard.
The review starts from the push itself, not from a pull request (a proposed change that someone approves before it merges). Direct pushes and force pushes to your production branch are checked too, which matters if you commit straight to main. If one push touches several Vercel projects, they are reviewed together.
What a finding gives you
Each finding records the project, environment, branch and commit, plus a summary, a suggested fix and a severity rating generated by AI. Vercel also adds its own check to each reviewed commit, and when a pull request exists it posts the findings as review comments. On public repositories, the details stay in the dashboard rather than appearing in public comments.
To fix a finding, you can click Fix with Vercel Agent. The agent proposes a change and opens a pull request, and nothing merges until you approve it. If you would rather use your own coding agent, run vercel security vulnerabilities in the Vercel command line tool, or copy a finding as Markdown or JSON and paste it in.
You can also write security policies in plain language, for all projects or specific ones. They give the reviewer context about how you build, such as which routes are meant to be public. Set a team policy in Agent settings, or a project policy in Security settings.
Cost and limits
Security reviews are free until November 10, 2026. After that, Vercel bills them at the Vercel Agent rate: the model provider’s token charges with no markup, plus $0.25 per million tokens. Fixes made with Vercel Agent are billed at the same rate. Because every push triggers a review, cost grows with how often you push and how much code changes.
Vercel says Security Review analyzes source code only. It is not a security audit and does no runtime or penetration testing, so a clean result does not mean your app is free of vulnerabilities. Severity ratings come from an AI model, not from a person who has watched your app run.
That makes it most useful as a second look at code a coding agent wrote. You have six days of free reviews, enough to enable it on one active project and see what it flags before billing starts. It covers a different risk from Vercel’s npm package blocking, which stops builds that install flagged dependencies. Security Review looks at the code you wrote.
End of article