Overview
AI Code Review Bot analyzes unified diffs and produces normalized findings (JSON or human-readable summary), with simple configuration for providers, policy, token budget, and VCS adapters (GitHub/GitLab/Bitbucket). By default it is safe: it uses a mock AI provider for deterministic runs and will not post PR/MR comments unless explicitly requested.
Key features
- Analyze unified diffs (files or PR/MR)
- Multiple output formats: JSON, human summary, and structured markdown
- YAML/JSON configuration with environment variable expansion
- Adapters for GitHub, GitLab, and Bitbucket
- Pluggable AI Providers: OpenAI, Gemini, Anthropic, Ollama, Mock
- Comment policies and secret redaction
- Security hardening with input sanitization and validation
- Performance optimization with intelligent chunking and caching
- Resource management with automatic cleanup
- Advanced token budgeting with compression
Privacy note: the default provider is mock to avoid external calls.
Enable real providers only when needed and configure tokens via environment variables.