At a Glance:
Kodus is an open-source AI code review tool that works in pull requests across GitHub, GitLab, Bitbucket, and Azure Repos, and supports any OpenAI-compatible LLM, including Claude, GPT, and Gemini.
Overview:
Kodus is an AI-powered code review assistant designed to integrate directly into the pull request workflow on major Git platforms. It allows development teams to automate code reviews using large language models while maintaining control over model selection and API costs, with no markup on LLM usage. The tool can be self-hosted or used as a cloud service, can learn from a project's architecture and standards, and supports custom review rules defined in plain language. It also provides a CLI for local and CI/CD pipeline runs.
Key Decision Points:
Model choice and cost control: Teams can bring their own API keys for any OpenAI-compatible model and pay providers directly, with no cost markup from the tool.
Deployment flexibility: Available as a self-hosted instance, a Kodus-hosted cloud service, or a mix, with support for self-hosted runners.
Git platform support: Natively integrates with GitHub, GitLab, Bitbucket, and Azure Repos for PR-based reviews.
Operational visibility: The Teams and Enterprise tiers include engineering metrics and a "Cockpit" for tracking technical debt and delivery metrics.
Community tier limitations: The free Community tier is limited to 10 custom rules and 3 active plugins and lacks metric dashboards, SSO, and RBAC.
Core Features:
Multi-model support: Works with Claude, GPT-5, Gemini, Llama, GLM, Kimi, and any provider exposing an OpenAI-compatible endpoint.
Custom review rules: Users can define plain-language rules to enforce specific review standards.
Kody learning and memory: The system adapts to a project's specific architecture, standards, and workflow.
CLI and CI/CD integration: Reviews can be triggered from the command line and within automated pipelines, beyond the PR interface.
Quality Radar: The Community edition surfaces unlimited "Quality Radar" issues to help assess code health.
Use Cases:
Development teams that want to automate PR code reviews using their preferred LLMs without paying a surcharge on API usage.
Teams with strict data privacy requirements that need a self-hosted code review tool with encrypted data handling and an opt-out telemetry mechanism.
Projects using multiple Git providers (GitHub, GitLab, Bitbucket, and Azure Repos) that require a single, consistent AI review process across all platforms.
Organizations that want to combine automated PR reviews with engineering metrics to monitor technical debt and delivery performance.
Open-Source Alternative Value:
Kodus provides an open-source core for AI-driven code review, enabling teams to self-host the solution and integrate their own language model providers via an OpenAI-compatible API. This approach lets users avoid markup on model usage and directly manage their costs with model providers. Its native support for multiple Git platforms and a CLI-run review mode offers flexibility beyond single-platform tools, while the self-hosted option allows for data handling according to internal policies.