Repository files navigation

A Streamlit-based chatbot powered by Retrieval-Augmented Generation (RAG) and OpenAI. Upload your PDFs and chat with them! This app leverages LangChain, FAISS, and OpenAI’s GPT models to extract and query document content with metadata-aware answers.

• 🔍Upload multiple PDFs and query across all of them
• 📄Metadata-rich answers with filename and page references
• 🧠 UsesLangChain + FAISS for semantic search
• 🤖Streamlit Chat UI for natural conversation
• 💾OpenAI API support with streaming responses

.├── .gitignore├── LICENSE├── README.md             # ← You're reading it├── app.py                # Main Streamlit app├── brain.py              # PDF parsing and vector index logic├── compare medium.gif    # Optional UI illustration├── requirements.txt      # Python dependencies└── thumbnail.webp        # Preview image

git clone https://github.com/co-dev0909/chatbot-using-rag-and-langchain.gitcd chatbot-using-rag-and-langchain

pip install -r requirements.txt

Create a.streamlit/secrets.toml file with:

OPENAI_API_KEY ="your-openai-key"

Or export it via environment variable:

export OPENAI_API_KEY="your-openai-key"

streamlit run app.py

1. Upload PDFs via the UI
2. Each PDF is parsed usingPyPDF2 and chunked via LangChain’sRecursiveCharacterTextSplitter
3. Chunks are embedded using OpenAI Embeddings
4. Stored in a FAISS vector store for semantic similarity search
5. Queries are matched to top PDF chunks and passed to ChatGPT with context
6. Answers includefile name andpage number metadata for citation

• Streamlit – UI framework
• LangChain – PDF chunking and retrieval
• FAISS – Vector search backend
• OpenAI GPT – LLM-based answer generation
• PyPDF2 – PDF parsing

"What are the main points from the introduction?"

Answer: The introduction highlights...(example.pdf, page 1)

This project is licensed under theMIT License.

Made with ❤️ byco-dev0909. Contributions welcome!