A chatbot that can answer questions about your personal documents using LangChain, ChromaDB, Cohere, and Streamlit.
- Upload and process multiple document types (PDF, DOCX, TXT)
- Interactive chat interface
- Semantic search using ChromaDB
- Powered by Cohere for accurate responses
- Clean and intuitive Streamlit UI
-Uploaded document
- Clone the repository
- Create and activate a virtual environment:
python3 -m venv venv source venv/bin/activate - Install dependencies:
pip install -r requirements.txt
- Copy
.env.exampleto.envand add your Cohere API key:cp .env.example .env
- Run the application:
streamlit run app.py
- Upload your documents using the sidebar
- Click "Process Documents" to index them
- Ask questions about your documents in the chat interface
- Get AI-powered responses based on your document content
- LangChain (>=0.1.0): For document processing and QA chain
- LangChain-Cohere (>=0.0.1): For Cohere integration
- ChromaDB (>=0.3.0): Vector store for document embeddings
- Cohere (>=4.37): For embeddings and chat completions
- Streamlit (>=1.22.0): Web interface
- Supporting libraries:
- pdfplumber (>=0.9.0): PDF processing
- python-docx (>=0.8.11): DOCX processing
- python-dotenv (>=1.0.0): Environment management