🔍 RAG Chatbot – Retrieval-Augmented Generation Chatbot with ChromaDB + HuggingFace + Gradio

Welcome to the RAG Chatbot, a powerful, fully open-source conversational AI system that can answer your questions using your own documents!

Built with:

• 🧠 LangChain for RAG pipeline logic
• 📚 ChromaDB for local document storage & retrieval
• ✨ Hugging Face Transformers for LLMs (like FLAN-T5)
• 💬 Gradio for a beautiful user interface
• ☁️ Hosted on Hugging Face Spaces

---

🚀 Features

• 📄 Upload documents (PDF, TXT, CSV, DOCX)
• 🔍 Ask questions and get context-aware answers
• 💾 Uses ChromaDB to store and search chunks
• 🧠 Uses FLAN-T5 or similar open-source models for generation
• 🌐 Works fully online — no local setup required
• 🔗 Shareable public URL with Hugging Face Spaces

---

📦 How It Works

1. Document Upload: Upload files from your device
2. Vector Store Creation: Text chunks are embedded and stored in ChromaDB
3. Query Answering: Your query is converted into a vector → similar chunks are retrieved
4. LLM Generation: Retrieved context is passed to the LLM (like flan-t5-base) to generate a response
5. Response Displayed in the Gradio UI

---

🧪 Try It Out

✅ Hosted version: https://huggingface.co/spaces/YasirUsman/rag-chatbot

---

🧰 Requirements

If you're running locally:

pip install -r requirements.txt
sudo apt-get install libmagic1

Or use this full setup in Colab:

!pip install -U langchain langchain-community huggingface_hub chromadb unstructured pypdf python-docx