Quodeq

Quodeq is an open-source, MIT-licensed tool that reviews a codebase with an AI model of the user's choice and grades it against selectable standards. It runs on the user's machine with no account, no Quodeq server and no telemetry, according to the project. The current release listed is v1.15.1.

What It Is

Quodeq is a code quality and security scanner. The user picks standards, the chosen model reads the code without changing it, and Quodeq reports a grade per dimension plus an overall grade. Every finding names the rule it breaks, the file and line, and the reason, and code done right is counted as compliant so grades reflect both sides. Built-in standards cover the six ISO 25010 quality dimensions (security, reliability, maintainability, performance, flexibility, usability), each graded 0 to 10, with CWE, OWASP ASVS, CISQ and WCAG 2.2 referenced in findings.

The Evaluate, Decide, Fix Loop

After an evaluation, users review each finding with its evidence and either send it to a fix plan or dismiss it. Fix plans list, per dimension, which files are affected, what is wrong and in what order to address it; users can fix issues themselves or hand the plan to an agent, then run the evaluation again. The dashboard offers an Overview with score history and trends, a code map showing where issues cluster, and project comparison.

Models and Standards

Quodeq works with AI CLIs the user already has (Claude Code, Codex, Gemini CLI, GitHub Copilot), cloud APIs (OpenRouter or any OpenAI-compatible endpoint), and local models via Ollama, llama.cpp or omlx, in which case code stays on the machine. Custom standards can be written in the dashboard, imported from a file, or drafted by an assistant from a named source; they are graded like built-in ones.

CI and Benchmarks

The quodeq review command posts findings as a review on GitHub pull requests and can write SARIF for code scanning; GitLab is not supported yet. Quodeq publishes OWASP Benchmark results for Java, which the vendor reports as one run on v1.12 and which have not been re-measured on newer versions. Desktop builds are available in beta as unsigned macOS and Windows apps.