🤖 ReviewX

中文文档 | English

Automated AI-powered code review for your pull requests. Works with any OpenAI-compatible API — just provide a base_url and API key. Also supports Anthropic Claude.

📸 Screenshot: ReviewX posting inline comments on a PR

---

✨ Features

• 🔄 GitHub Action — Auto-reviews PRs on open/sync
• 🤖 Universal Provider Support — Any OpenAI-compatible API (MiMo, OpenAI, DeepSeek, Qwen, Kimi, etc.) + Anthropic Claude
• 📍 Inline Comments — Reviews specific lines in your diff
• 📊 Review Scoring — 0-100 score with severity breakdown
• 🔍 Smart Detection — Bugs, security, performance, style
• ⚙️ Configurable — Exclude patterns, severity filters, custom rules
• 🖥️ CLI Mode — Review files and diffs from terminal
• 🌐 Webhook Server — Standalone mode with GitHub webhooks
• 📈 Dashboard — Flask web UI for review history & stats

---

🚀 Quick Start

As a GitHub Action (Recommended)

1. Add secrets to your repository:

   • MIMO_API_KEY — Your MiMo/OpenAI API key

2. Create .github/workflows/review.yml:

name: AI Code Review
on:
  pull_request:
    types: [opened, synchronize, reopened]

permissions:
  contents: read
  pull-requests: write

jobs:
  review:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - uses: peppa486/reviewx@main
        with:
          github-token: ${{ secrets.GITHUB_TOKEN }}
          ai-provider: openai
          openai-api-key: ${{ secrets.MIMO_API_KEY }}
          openai-base-url: https://token-plan-sgp.xiaomimimo.com/v1
          openai-model: mimo-v2.5-pro

3. Open a PR and watch the magic! ✨

CLI Mode

# Install
pip install -r requirements.txt

# Review a PR
python cli.py review-pr --repo owner/repo --pr 42 --post-comments

# Review a single file
python cli.py review-file src/main.py

# Review a git diff
git diff HEAD~1 | python cli.py review-diff

# Start webhook server
python cli.py serve

Standalone Webhook Server

# Copy and edit config
cp config.yaml.example config.yaml

# Start server
python bot.py

Configure a GitHub webhook pointing to https://your-server:8080/webhook.

---

🔧 Configuration

Environment Variables

Variable	Description	Default
AI_REVIEW_PROVIDER	openai or anthropic	openai
OPENAI_API_KEY	API key for OpenAI-compatible providers	—
OPENAI_BASE_URL	API base URL	https://token-plan-sgp.xiaomimimo.com/v1
OPENAI_MODEL	Model name	mimo-v2.5-pro
ANTHROPIC_API_KEY	Anthropic API key	—
GITHUB_TOKEN	GitHub personal access token	—

Config File

See config.yaml.example for full options including:

• Severity level filters
• File exclusion patterns
• Review categories
• Server settings

---

🤖 Supported AI Providers

ReviewX works with any OpenAI-compatible API — just provide a base_url and API key. No need for provider-specific code.

OpenAI-Compatible Providers (use openai provider)

Any service that implements the OpenAI chat completions API works out of the box:

• Xiaomi MiMo — MiMo-V2.5 Pro (https://token-plan-sgp.xiaomimimo.com/v1)
• OpenAI — GPT-5.5 (https://api.openai.com/v1)
• DeepSeek — DeepSeek V4 Pro, DeepSeek V4 Flash (https://api.deepseek.com)
• Moonshot / Kimi — Kimi K2.6 (https://api.moonshot.cn/v1)
• Zhipu GLM — GLM-5.1 (https://open.bigmodel.cn/api/paas/v4)
• Alibaba Qwen — Qwen 3.6 Max (https://dashscope.aliyuncs.com/compatible-mode/v1)
• ByteDance Doubao — Doubao Seed 2.0 Pro (https://ark.cn-beijing.volces.com/api/v3)
• xAI — Grok 4.3 (https://api.x.ai/v1)
• MiniMax — M2.7 (https://api.minimaxi.com/v1)
• Mistral — Medium 3.5 (https://api.mistral.ai/v1)
• Any self-hosted — Ollama, vLLM, LiteLLM, etc.

Anthropic Claude (uses anthropic provider)

Claude uses the Anthropic API format (not OpenAI-compatible), so it needs a separate provider config:

• Claude Opus 4.7 (claude-opus-4-7)
• Set ai-provider: anthropic and provide ANTHROPIC_API_KEY

---

📊 Dashboard

The optional Flask dashboard shows review history and statistics:

python dashboard/app.py

Visit http://localhost:5000 to see:

• Total reviews and average scores
• Issue severity breakdown
• Recent review history

---

📁 Project Structure

reviewx/
├── action.yml              # GitHub Action definition
├── bot.py                  # Webhook server entry point
├── cli.py                  # CLI entry point
├── config.yaml.example     # Configuration template
├── src/
│   ├── reviewer.py         # Core review orchestration
│   ├── providers/
│   │   ├── base.py         # Provider interface & prompts
│   │   ├── openai_provider.py
│   │   └── anthropic_provider.py
│   ├── github_client.py    # GitHub API client
│   ├── diff_parser.py      # Git diff parser
│   ├── comment_formatter.py
│   └── config.py           # Configuration management
├── dashboard/              # Flask web dashboard
├── tests/                  # Test suite
└── .github/workflows/      # Example workflow

---

🧪 Testing

pip install pytest
pytest tests/ -v

---

📄 License

MIT License — see LICENSE for details.

---

🙏 Credits

Built with ❤️ by peppa486

Powered by Xiaomi MiMo and other leading AI models.