# Verity

> Independent code reviewer for AI coding agents. Verity reviews what your AI agent just did against a Standard you control — catching drift, risky changes, and scope creep before they land.

Verity installs into Claude Code via a CLI. It runs at key moments (agent stop, pre-commit, pre-push), scores changes against a project-specific Standard, and maintains a memory of what the project has learned.

## Product
- [Verity](/): What Verity is and how it fits into an AI coding workflow
- [Live demo report](/demo): Example review output from a real run
- [Changelog](/changelog): Recent product updates
- [Tolaria](/tolaria): Companion landing page

## Docs
- [Introduction](/docs): Overview of Verity and its concepts
- [Quickstart](/docs/quickstart): Install the CLI and run /verity-setup
- [How it works](/docs/how-it-works): The review loop end-to-end
- [Moments](/docs/moments): Stop, pre-commit, and pre-push gates
- [The Standard](/docs/standard): Project-specific review rules
- [Memory](/docs/memory): The knowledge base Verity builds over time
- [Cost & usage](/docs/cost): Token cost and usage reporting
- [Security & privacy](/docs/security): Zero-repo-access, in-memory diffs
- [CLI & commands](/docs/cli): verity CLI + Claude Code skills reference
- [Troubleshooting](/docs/troubleshooting): Common issues and fixes