Self-hosted AI code review bot for GitHub PRs.

MIT-licensed alternative to CodeRabbit, Qodo Merge, and Copilot Code Review. Bring your own LLM — OpenAI, Anthropic, DeepSeek, or local Ollama.

git clone https://github.com/kunci115/open-review
cd open-review
cp .env.example .env   # fill in LLM_API_KEY + GITHUB_TOKEN
docker compose up -d

Create a webhook on your repo:

• URL: https://your-server:4000/webhook
• Events: Pull requests
• Content type: application/json

Or use the public endpoint via ngrok for testing:

ngrok http 4000

Open a PR → OpenReview posts a review within 2-4 minutes.

DeepSeek (recommended — cheap + good):

LLM_BASE_URL=https://api.deepseek.com/v1
LLM_MODEL=deepseek-chat

OpenAI:

LLM_BASE_URL=https://api.openai.com/v1
LLM_MODEL=gpt-4o

Anthropic (via compatible proxy):

LLM_BASE_URL=https://api.anthropic.com/v1
LLM_MODEL=claude-sonnet-4

Ollama (local, free):

LLM_BASE_URL=http://localhost:11434/v1
LLM_MODEL=qwen3-coder

• 🔴 Security: SQL injection, XSS, hardcoded secrets, unsafe deserialization
• 🟡 Bugs: null refs, race conditions, incorrect error handling
• 🟠 Performance: N+1 queries, missing indexes, inefficient loops
• 🔵 Patterns: anti-patterns, broken conventions, missing validation

• GitLab + Bitbucket support
• openreview.yaml config file (path-specific rules)
• PR summarization (/describe-style)
• Code suggestions with fix snippets
• Review dashboard UI
• Slack/Discord notifications

OpenReview is MIT licensed, self-hosted, and will stay that way forever.

If it saves your team hours of code review, consider supporting:

Built with ❤️ by Rino • MIT • Self-hosted • No vendor lock-in