NotebookLM-Lite is a note-taking application inspired by Google's NotebookLM, aiming to fully replicate features such as document Q&A, podcast generation, and presentation creation. It is currently under development. For a simple Chinese podcast generation feature, please refer to the v0.1 branch.
- Unified LLM Interface: Seamlessly switch between OpenAI and Anthropic models.
- Neural Audio Synthesis: High-quality, multi-speaker podcast generation using CosyVoice (via Dashscope).
- PDF Intelligence: Robust parsing and semantic understanding of uploaded documents.
- Interactive Player: Integrated audio player with playback controls and download capability.
- Frontend: React, Vite, TailwindCSS, Framer Motion (animations)
- Backend: Python, FastAPI, Uvicorn
- AI/ML: OpenAI API, Anthropic API, Dashscope (TTS)
- Python 3.10+
- Node.js 18+
- FFmpeg (Must be installed and added to system PATH)
Navigate to the project root:
# Install Python dependencies
pip install -r requirements.txt
# Start the API server
python backend/main.pyThe backend will run at http://localhost:8000.
Open a new terminal and navigate to the frontend directory:
cd frontend
# Install Node dependencies
npm install
# Start the development server
npm run devThe UI will be available at http://localhost:5173.
- Launch: Open
http://localhost:5173in your browser. - Configure:
- Select your LLM Provider (OpenAI Compatible or Anthropic).
- Enter your API Key and Model Name (e.g.,
gpt-4o). - Enter your Dashscope API Key for TTS service.
- Upload: Drag & drop a PDF file into the "Data Source" area.
- Generate: Click INITIALIZE SEQUENCE. The system will process the text and synthesize audio.
- Listen: Use the built-in player to listen to your podcast or download it.
Apache 2.0