🤖 RoboChat

RoboChat is an AI-powered assistant designed for electronics, robotics, and IoT learners.
It combines local knowledge retrieval, web search integration, and Hugging Face inference models to provide:

• Technical explanations for microcontrollers, sensors, and robotics modules.
• Real-time store lookup and pricing information for components.
• Interactive tutor mode for step-by-step project guidance.
• Full deployment with backend + frontend, making it usable as a standalone web app.

---

🚀 Features

• LLM Integration: Powered by Hugging Face Inference (openai/gpt-oss-120b:fireworks-ai).
• Graph-based Tool Use: Runs reasoning pipelines with retrieval + web tools.
• Frontend: Simple, lightweight interface (HTML, CSS, JS only, no frameworks).
• Backend: Python FastAPI backend to handle inference calls.
• Deployment Ready: Works on local setup or can be hosted on cloud platforms.

---

📂 Project Structure

RoboChat/
    │── backend/ # Agent, Tools, FastAPI backend, connects to Hugging Face inference
    │── frontend/ # Simple HTML/CSS/JS interface
    │── notebook.ipynb # Jupyter notebooks for experiments & demos
    │── README.md # Project documentation

---

🛠️ Setup & Installation

1️⃣ Clone Repository

git clone https://github.com/Seif-Sameh/RoboChat.git
cd robochat

2️⃣ Backend Setup

cd backend
pip install -r requirements.txt
uvicorn main:app --reload --port 8000                                                                  

3️⃣ Frontend Setup

Frontend is a simple static app (no frameworks). Open frontend/index.html in a browser, or serve via a local web server:

cd frontend
python -m http.server 5500

Now visit: http://localhost:8000

---

📒 Notebooks

Inside notebooks/ you’ll find:

• robochat_demo.ipynb → Example of querying RoboChat with LangChain graph.
• Cells showing local Q&A and web-assisted queries.

---

🧪 Example Usage

Inside Jupyter notebooks, you can test RoboChat’s reasoning:

result = bot.graph.invoke({"messages": [HumanMessage(content="What is Arduino UNO?")]})
print(result['messages'][-1].content)

---

📦 Deployment

This project is designed for full-stack deployment:

• Backend → FastAPI with Hugging Face inference.
• Frontend → Simple HTML/CSS/JS UI for interaction.
• Optional Hosting → Deploy on platforms like Vercel (frontend) + Render/Heroku (backend).

---

🔮 Future Plans

• Expand knowledge base for advanced robotics topics.
• Multi-modal support (images + circuit diagrams).

---

💡 Credits

• Hugging Face Inference API
• LangChain for orchestration
• OpenAI OSS Models (gpt-oss-20b, gpt-oss-120b)