Quickstart

From install to your first findings, on your own machine. No account, nothing to sign up for.

1. Install

pipx install quodeq

It needs Python 3.12 or newer. On macOS you can also use Homebrew, brew install quodeq/tap/quodeq, or download the macOS app from the latest release.

2. Have a model ready

quodeq uses an AI model to read your code. Use one you already have:

• An AI CLI you are signed in to: Claude Code, Codex, Gemini CLI or GitHub Copilot. Its sign-in and billing apply.
• A cloud API key: OpenRouter, or any OpenAI-compatible endpoint.
• A local model: Ollama, llama.cpp or omlx. The code never leaves your machine.

If you already use Claude Code or Codex, pick that and there is nothing to set up. OpenRouter is the cheapest way to start with a cloud model, and Ollama keeps everything on your machine. You choose the model in Settings. Setup for each one is in AI Providers.

3. Run your first evaluation

quodeq

This opens the dashboard. Add a repository (a git URL or a local folder) and press scan and run. The run streams live, and you can watch findings appear as the model reads the code.

Prefer the terminal? quodeq evaluate /path/to/repo runs the same evaluation without the UI.

4. Read the results

A first run reads every file: 2,803 Java files took 32 minutes with Claude Sonnet. Later runs reread only what changed. When the run finishes, the Overview shows a grade for each quality dimension. Violations lists every finding with its file, line, reason and the code. Each dimension has a fix plan you can follow yourself or give to a coding agent. Run quodeq again afterwards to check the problems are gone.

Where to go next

Check every pull request with Command Line & CI. See how findings become a grade in Why This Grade.

Watch it

An overview of quodeq, recorded on an earlier version, so some screens look different from 1.13. The video loads from youtube-nocookie.com only when you press play. More videos