A tool that transforms short, vague user prompts into well-structured, optimized system instructions for Large Language Models. Input a basic idea, and get a production-ready AI prompt to use in your workflow.
The project is built on a microservices architecture, keeping the frontend, API gateway, and AI generation service completely independent.
- Frontend: Next.js 14 (TypeScript) styled with Tailwind CSS, featuring smooth animations powered by Framer Motion. Includes a dark theme and a typewriter effect for AI text generation.
- API Gateway: Express (TypeScript) handles incoming requests, request validation, and rate-limiting.
- AI Service: Python (FastAPI) handles the connection to LLM provider APIs.
- Infrastructure: Docker and Docker Compose for containerization and quick local orchestration.
Note: If you do not have any LLM API keys configured, the project automatically switches to a local Fallback Engine that builds prompts using pre-defined templates entirely offline.
blueprint-ai/
├── frontend/ # Next.js user interface
│ └── src/
│ ├── app/ # Routing and pages
│ ├── components/# UI elements (Forms, chat boxes, buttons)
│ └── lib/ # API client for backend requests
├── backend/ # Express API Gateway (Validation & Security)
│ └── src/
│ ├── routes/ # API endpoints
│ └── controllers/# Gateway route logic
├── ai-service/ # Isolated Python service
│ └── app/
│ └── services/ # Core prompt engineering logic
├── docker-compose.yml # Single-command container deployment setup
└── .env.example # Environment configuration template
- Start the AI Service (Python):
cd ai-service && source venv/bin/activate && uvicorn app.main:app --reload --port 8000
- Start the API Gateway (Express):
cd backend && npm run dev
- Start the Frontend (Next.js):
cd frontend && npm run dev
Once all services are running, open your browser and navigate to http://localhost:3000.
To skip installing Node.js and Python dependencies directly on your machine, you can run the entire system using Docker. Make sure the Docker daemon is running, then execute the following command in the root folder:
docker compose up --build
Docker will download the images, spin up the containers, and map the ports automatically. The application will be available at http://localhost:3000.
Key environment variables you can customize in your .env file:
* PORT — API Gateway port (defaults to 4000).
* FRONTEND_URL — Frontend URL used for CORS configurations (http://localhost:3000).
* OPENAI_API_KEY — Your LLM API key (optional; if left blank, the app uses local templates).
* OPENAI_MODEL — The specific AI model handling requests (e.g., gpt-4o or your chosen DeepSeek model).
* MAX_TOKENS — Response token limit for prompt generation (set to 1500).
* NEXT_PUBLIC_API_URL — URL pointing the frontend to the gateway (http://localhost:4000/api).
This project is open-source and licensed under the MIT License. Feel free to fork, modify, and use it as you see fit.