Kodus
AI code review software with a self-hosted AGPL core. Use Claude, GPT, Gemini, Llama or self-hosted models through OpenAI-compatible endpoints.
Kodus is an AI code review tool for engineering teams that want automated pull request feedback while choosing where the reviewer runs and which model it uses. Its open source core uses the AGPL license and can be self-hosted with Docker Compose or on Kubernetes. Kodus also offers a hosted cloud service that manages the infrastructure.
You bring your own model credentials in either setup. It works with Claude, GPT, Gemini and Llama, including self-hosted models through OpenAI-compatible endpoints. Teams can choose a model per repository and a fallback model. Self-hosting controls where Kodus runs; the selected model endpoint determines where AI processing happens. Kodus says it doesn't store source code or use customer data to train models.
The reviewer, Kody, checks changes against surrounding callers, tests and code in linked repositories. It posts inline feedback in pull requests through GitHub, GitLab, Bitbucket, Azure DevOps and Forgejo/Gitea. Reviews can also run locally through its CLI or in CI pipelines.
Teams can write review rules in plain language or import conventions from .cursorrules, CLAUDE.md and AGENTS.md. Kody can check changes against requirements from Jira, Linear and Notion. Developers can discuss flagged issues inside the pull request, and Kody can save their corrections as memory for future reviews. Unimplemented suggestions from closed pull requests can become tracked issues, which Kody resolves when it detects a fix in a later pull request.
Self-hosted instances send a daily anonymous usage heartbeat by default; it can be disabled with KODUS_TELEMETRY_DISABLED.
