# Verity

> A pair programmer to your coding agent. Verity reviews what your coding agent writes while it works, at each turn, commit and push, against a Standard your team controls, so problems are fixed before a PR exists.

Verity installs into Claude Code with a CLI: `npm install -g @codacy/verity-cli && verity init`. It keeps a Markdown memory of each project's decisions and gotchas, and meters what your agents spend. Built by Codacy.

Each page below links to a Markdown copy of its text. [llms-full.txt](/llms-full.txt) holds all of them in one file.

## Product
- [Verity](/index.md): What Verity is and how it fits into an AI coding workflow
- [Why the PR is too late](/why.md): Why pull requests are piling up, with sources, and why review has to move earlier
- [Benchmark](/benchmark.md): How Verity was tested, the results, and the method
- [Review gates](/product/gates.md): The review at each agent turn, commit and push, what it checks and what blocks
- [Formal verification](/product/formal.md): How Verity models one risky function in Lean at push, searches it for a failing order of events, and replays the failure on the real code before it can block
- [Your Standard](/product/standard.md): One versioned set of rules for every agent, built on public review packs
- [Project memory](/product/memory.md): The Markdown knowledge base of decisions and gotchas that agents and the reviewer read
- [Cost](/product/cost.md): Token spend per model, session, repository and organisation
- [Pricing](/pricing.md): The Free, Pro, Fleet and Enterprise plans, with their prices and limits

## Docs
- [Introduction](/docs.md): Overview of Verity and its concepts
- [Quickstart](/docs/quickstart.md): Install the CLI and set Verity up in a project
- [Authentication](/docs/authentication.md): How Verity signs you in
- [How it works](/docs/how-it-works.md): The review loop from end to end
- [Moments](/docs/moments.md): When Verity reviews: each turn, before commit and before push
- [Formal verification](/docs/formal.md): Lean models of one risky function at push; counterexamples replayed on the real code; which pushes qualify; languages
- [Findings and verdicts](/docs/findings.md): What a finding is made of and how to answer one
- [The Standard](/docs/standard.md): Project-specific review rules
- [Memory](/docs/memory.md): The knowledge base Verity builds over time
- [Org knowledge](/docs/org-knowledge.md): Knowledge shared across an organisation's repositories
- [Token telemetry and cost](/docs/cost.md): What your agents spend in tokens and estimated USD, from their own telemetry
- [Dashboard](/docs/dashboard.md): What the dashboard shows
- [Security and privacy](/docs/security.md): What Verity sees of your code and what it keeps
- [CLI and commands](/docs/cli.md): The verity CLI and Claude Code skills reference
- [Files](/docs/files.md): The files Verity writes in your repository
- [Troubleshooting](/docs/troubleshooting.md): Common issues and fixes

## Optional
- [Rule catalog](/catalog): Every review pack and its rules, with examples (loads in the browser)
- [Changelog](/changelog): Recent product updates (loads in the browser)
- [Live demo report](/demo): Example review output from a real run