kodustech/kodus-ai
AI Code Review with Full Control Over Model Choice and Costs.
What it solves
Kodus AI provides an automated AI-powered code review system that integrates directly into the Git workflow. It helps engineering teams reduce technical debt, enforce architecture and security standards, and ensure high review quality without requiring manual repetition of the same feedback by human reviewers.
How it works
Kodus operates as a monorepo containing an API, a background worker for review execution, and a webhook service for Git providers (GitHub, GitLab, Bitbucket, and Azure Repos). It uses a "Bring Your Own Key" (BYOK) model, allowing teams to connect their own LLM providers (such as Claude, GPT-5, Gemini, or Llama) to analyze pull requests. Teams can define "Kody Rules" in plain language to guide the AI's review process based on specific project standards.
Who it’s for
Software engineering teams and organizations that want to automate code reviews, track technical debt, and maintain consistent coding standards across multiple repositories.
Highlights
- Model Agnostic: Supports a wide range of LLMs via OpenAI-compatible endpoints.
- Custom Review Rules: Define plain-language instructions to enforce architecture and security expectations.
- Native Git Integration: Works directly within pull requests across major Git providers.
- Comprehensive Tooling: Includes a CLI for local/CI-CD reviews, a dashboard for token usage and delivery metrics, and a system to track unimplemented suggestions as issues.
- Privacy-Focused: Supports self-hosting and ensures source code is not used for model training.