Browse Code
Getzilla is an open source AI code review and quality gate for vibe coding. Your AI coding agent writes the code on its own branch; Getzilla checks it before it stays. Tests really run, a second AI reviews the change, the author can never approve itself, and every task gets a pull request with saved evidence: plan, test results and reviews.
Use it to make AI generated code safer and production ready: catch problems before merge, keep AI technical debt down, no unchecked pushes to main. Lessons go to mistakes.md, so the next session does not repeat them.
Self-hosted: runs locally in your repo with Qwen Code, Codex, Claude Code, Gemini CLI, Copilot CLI or Grok Build. Free, MIT license. Mirror of github.com/Dimkox/Getzilla. Site: getzilla.app
Features
- Open source AI code review tool: a second AI reviews every change
- AI code review agent that never lets the author approve its own code
- Quality gate for vibe coding: AI agent code stays only after checks pass
- Test automation with AI agents: tests really run, results saved per task
- Every task gets its own branch and pull request, no pushes to main
- Review AI generated code with saved evidence: plan, tests and reviews
- Vibe coding security: risky changes get extra checks and human sign-off
- Take a vibe coded prototype to production with traceable checks
- Keep AI technical debt down: lessons in mistakes.md, decisions in decisions.md
- Stops instead of guessing when something is unclear
- Self-hosted AI code review: runs locally in your repo, your own models
- Works with Qwen Code, Codex, Claude Code, Gemini CLI, Copilot CLI, Grok Build
- Pinned installer with SHA-256 check; read-only demo tour
- Free AI code review tool, MIT license; optional EU AI Act evidence layer
Project Samples
