Ai Review Pipeline

ai-review-pipeline

AI-powered code quality pipeline CLI — Review + Test + Report in one command. Add `--fix` for auto-fix loop.

**English | [中文](./README.zh-CN.md)**

---

Features

• Zero config — Built-in free AI model, npx ai-review-pipeline works out of the box with no API Key
• Streaming output — Real-time token-by-token response, no more staring at a blank screen
• Large diff chunking — Auto-splits oversized diffs by file for complete coverage
• Zero dependencies — No required deps, instant npx execution
• Multi-provider — OpenAI / DeepSeek / Claude / Qwen / Gemini / SiliconFlow / Ollama, auto-detected
• Unified pipeline — Default: Review + Test + Report (read-only). --fix enables auto-fix loop
• Flexible targets — Files, folders, comma-separated multi-targets
• --full mode — Review entire file content without git changes
• HTML reports — Score + issue list + fix suggestions, attachable to PRs
• JSONC config — .ai-pipeline.json with comments support, shared across team
• CI-ready — --json output + exit codes for GitHub Actions / GitLab CI
• Bilingual prompts — --lang en switches both output AND AI prompt to English
• Multi-language code — TypeScript / JavaScript / Vue / Python / Go / Rust / Java / Swift / PHP / Kotlin

Quick Start

# Zero config — just run it (built-in free model)
npx ai-review-pipeline

# Auto-fix pipeline
npx ai-review-pipeline --fix

💡 No API Key needed to try the full pipeline. The built-in free model has rate limits — configure your own Key for a faster and more stable experience.

Configure your own API Key (optional, recommended)

# Pick any provider and add to .env.local

# DeepSeek (affordable & capable)
echo 'DEEPSEEK_API_KEY=sk-xxx' >> .env.local

# OpenAI
echo 'OPENAI_API_KEY=sk-xxx' >> .env.local

# Claude
echo 'ANTHROPIC_API_KEY=sk-ant-xxx' >> .env.local

# Qwen (Alibaba Cloud)
echo 'DASHSCOPE_API_KEY=sk