> For the complete documentation index, see [llms.txt](https://docs.visdom.virtuslab.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.visdom.virtuslab.com/visdom-code-review/visdom-code-review.md).

# Visdom Code Review

Multi-layered review for teams shipping AI-generated code.

Your CI says the code is fine. The tests pass. The AI wrote the tests too.

That is the problem this component exists for. Review was calibrated on the assumption that code and tests came from different heads, and that badly reasoned code tended to look badly written. Neither holds for generated code, so the signals a reviewer relies on stop arriving.

Visdom Code Review is a **review process**, not a service. It runs inside your CI, against your model provider, behind your network boundary, and the policy it enforces lives in your repository as a versioned file rather than in a vendor's console.

## What it is not

It is not a linter with a language model bolted on, and it is not a replacement for a reviewer. It classifies changes, spends effort proportional to risk, and hands a person the few per cent that genuinely need judgement, with the reasoning attached.

## How it is delivered

The reference implementation is open source and the pipeline runs in your own infrastructure. A deployment is an engagement, not a subscription: the pipeline is configured for your stack, the review lenses are tuned to your conventions, and your platform team owns and operates it afterwards. Code does not leave your network, and there are no per-seat fees, only the model costs you already pay.