AI Prompt Factory - Automated High-Quality Prompt Suite Generation
After the system runs, various roles required by agents will be created, responsible for different functions!
It's not a single prompt word, it's a complete set of agent prompt word engineering!!!!
Prompt Factory is a multi-agent collaborative prompt generation system. Input your requirements, and the system will automatically:
- Analyze - Analyzer breaks down requirements, designs system architecture and roles
- Generate - Generator creates professional prompts for each role
- Review - Reviewer evaluates prompt quality, scores and suggests improvements
- Optimize - Optimizer refines prompts based on review until they pass
Output is a structured set of prompt files ready for use in AI applications.
- 🚀 Multi-agent pipeline automation
- 🔄 Review-optimize iteration loop for quality assurance
- 💾 Checkpoint recovery, resume after interruption
- 🌍 Bilingual interface (Chinese/English)
- 📁 Real-time saving to Markdown files
- Python 3.9+
- Node.js 18+
- pnpm (recommended) or npm
Backend:
cd factory/server
pip install -r requirements.txtFrontend:
cd factory/client
pnpm installOption 1: Using scripts
# Start backend
start_server.bat
# Start frontend (new terminal)
start_client.batOption 2: Manual
# Backend
cd factory/server
python run.py
# Frontend
cd factory/client
npm run devOpen browser at http://localhost:5173
On first use, click the settings button (top right) to configure:
- API Key: Your LLM API key
- Base URL: API endpoint (OpenAI-compatible format)
- Model: Select the model to use
- Select prompt type (Programmer, AI Image, Customer Service, etc.)
- Fill in or modify the requirement description
- Click "Start Generate"
- Wait for the agent pipeline to complete
- Check generated prompt files in the results directory
factory/
├── client/ # Frontend React app
├── server/ # Backend FastAPI service
│ ├── app/
│ │ ├── services/ # Core services (pipeline, llm, storage)
│ │ ├── routes/ # API routes
│ │ └── prompts/ # Agent prompt templates
│ └── results/ # Generated output directory
└── README.md
AI 提示词工厂 - 自动化生成高质量提示词套件
系统运行后会创建各种agent 需要的角色,负责不同的职能!
不是单一的提示词,是一整套agent 提示词工程!!!
Prompt Factory 是一个基于多 Agent 协作的提示词生成系统。输入你的需求描述,系统会自动:
- 分析需求 - Analyzer 拆解需求,设计系统架构和角色分工
- 生成提示词 - Generator 为每个角色生成专业提示词
- 质量审核 - Reviewer 评估提示词质量,打分并提出改进建议
- 迭代优化 - Optimizer 根据审核结果优化提示词,直到达标
最终输出一套结构化的提示词文件,可直接用于各类 AI 应用。
- 🚀 多 Agent 流水线自动化生成
- 🔄 审核-优化迭代循环,确保质量
- 💾 断点恢复,支持中断后继续
- 🌍 中英双语界面
- 📁 结果实时保存为 Markdown 文件
- Python 3.9+
- Node.js 18+
- pnpm(推荐)或 npm
后端:
cd factory/server
pip install -r requirements.txt前端:
cd factory/client
pnpm install方式一:使用启动脚本
# 启动后端
start_server.bat
# 启动前端(新终端)
start_client.bat方式二:手动启动
# 后端
cd factory/server
python run.py
# 前端
cd factory/client
npm run dev打开浏览器访问 http://localhost:5173
首次使用时,点击右上角设置按钮,配置:
- API Key:你的 LLM API 密钥
- Base URL:API 地址(默认兼容 OpenAI 格式)
- 模型:选择要使用的模型
- 选择提示词类型(程序员助手、AI绘图、客服系统等)
- 填写或修改需求描述
- 点击「开始生成」
- 等待 Agent 流水线执行完成
- 在结果目录查看生成的提示词文件
factory/
├── client/ # 前端 React 应用
├── server/ # 后端 FastAPI 服务
│ ├── app/
│ │ ├── services/ # 核心服务(pipeline, llm, storage)
│ │ ├── routes/ # API 路由
│ │ └── prompts/ # Agent 提示词模板
│ └── results/ # 生成结果输出目录
└── README.md