Ask Questions. Get Answers. All from Your Own Documents—Securely and Instantly.
DocuChat AI is a private, smart, and beautiful document Q&A web app.
Upload any PDF, DOCX, or TXT file. Instantly ask questions and get clear, context-aware answers—powered entirely by local AI.
No data leaves your device. Your privacy, your answers.
- 🧠 Chat With Your Documents: Upload resumes, reports, academic papers, notes, or contracts—ask anything and get relevant answers in seconds.
- 🔒 100% Local & Private: No cloud, no API calls. Everything runs on your machine with HuggingFace Transformers and FAISS.
- ⚡ Ultra-fast, Lightweight Models: Uses TinyLlama or Phi-2 for instant answers, even on laptops.
- 📄 Semantic Search: Finds the best context in your document using state-of-the-art embeddings.
- 🎨 Professional UI: Modern Bootstrap, animated loader, collapsible context, and mobile responsive.
- 🖼️ Stunning Look: Landing page, upload flow, and Q&A all designed for clarity and trust.
- Flask (Python 3.8+) – backend and routing
- Bootstrap 4/5 & FontAwesome – modern responsive UI
- HuggingFace Transformers – local LLM (TinyLlama, Phi-2, etc.)
- SentenceTransformers – fast local embeddings
- FAISS – vector similarity search for semantic retrieval
- PyPDF2, python-docx – robust text extraction
- Upload a document (PDF, DOCX, or TXT)
- Text is extracted, chunked, and embedded
- Semantic search finds the most relevant parts for your question
- Local LLM answers your question using both your context and the original question
- Results displayed instantly, with context you can review and trust
Clone and run in under 3 minutes:
git clone https://github.com/MuhammadSaqib786/docuchat-ai.git
cd docuchat-ai
python -m venv venv
source venv/bin/activate # On Windows: venv\Scripts\activate
pip install -r requirements.txt
python app.py