JupyterWhisper transforms your Jupyter notebook environment by seamlessly integrating Claude AI capabilities. This extension enables natural chat interactions, intelligent code execution, and voice command features to enhance your notebook workflow.
- 🤖 Native integration with Claude 3.5 Sonnet
- 🎯 Intelligent code execution and cell management
- 🔍 Advanced search capabilities powered by Perplexity AI
- 🎙️ Voice command support using OpenAI Whisper
- 📝 Context-aware text processing and formatting
- 💬 Comprehensive chat history management
- ⚡ Real-time streaming responses
pip install jupyter_whisper- Python 3.7+
- JupyterLab 4.0+ (important: this extension is designed for JupyterLab, not classic Notebook)
- Jupyter Notebook 7.0+ (if using Notebook instead of Lab)
- Required API keys:
- Anthropic API key (for Claude integration)
- OpenAI API key (optional, for voice features)
- Perplexity API key (for advanced search capabilities)
- Install JupyterLab if you haven't already:
pip install jupyterlab>=4.0.0- Install Jupyter Whisper:
pip install jupyter_whisper- Start JupyterLab:
jupyter labJupyter Whisper runs a local FastAPI server (on port 5000) to handle features like audio transcription and text processing. The server is shared between notebooks for efficiency.
Important Notes:
- The server persists between notebook sessions
- Configuration changes (like API keys) only take effect when the server restarts
- You'll be notified if you're using an older server version
To manually refresh the server and apply new configurations:
from jupyter_whisper import refresh_jupyter_whisper
refresh_jupyter_whisper() # Warning: affects all active notebooksWhen to refresh:
- After updating API keys
- After upgrading the package
- If you encounter configuration issues
Note: Refreshing the server will impact all notebooks currently using it. You may need to restart kernels in affected notebooks.
JupyterWhisper is specifically designed and tested for JupyterLab 4.0+. While it may work in classic Jupyter Notebook (7.0+), we recommend using JupyterLab for the best experience and full feature support.
Key compatibility notes:
- Voice features require a modern browser
- WebSocket support is required for real-time streaming
- Some features may require JupyterLab extensions to be enabled
- Port 5000 must be available for the local server
The easiest way to configure Jupyter Whisper is through the interactive setup interface:
import jupyter_whisperThis will open an interactive UI with tabs for:
- API Keys configuration
- Model selection
- System prompt customization
You can also configure settings programmatically:
from jupyter_whisper.config import get_config_manager
config = get_config_manager()
# Set API keys
config.set_api_key('ANTHROPIC_API_KEY', 'your-key-here')
config.set_api_key('OPENAI_API_KEY', 'your-key-here') # Optional for voice
config.set_api_key('PERPLEXITY_API_KEY', 'your-key-here') # For search
# Change the model
config.set_model('claude-3-5-sonnet-20241022')
# Update system prompt
config.set_system_prompt("Your custom system prompt here")
# Set other preferences
config.set_config_value('SKIP_SETUP_POPUP', True)Available models:
- claude-3-5-sonnet-20241022
- claude-3-5-haiku-20241022
- claude-3-opus-20240229
- claude-3-sonnet-20240229
- claude-3-haiku-20240307
Interact with the AI using the %%user magic command:
%%user
How do I read a CSV file using pandas?Access web information directly within your notebook:
from jupyter_whisper import search_online
style = "Be precise and concise"
question = "What's new in Python 3.12?"
search_online(style, question)Leverage voice input capabilities:
- Control recording with keyboard shortcuts
- Automatic speech-to-text conversion
- Seamless chat interface integration
%%user [index]- Initiate a user message%%user [index]:set- Replace user message at given index%%assistant [index]- Include assistant response%%assistant [index]:set- Replace assistant message at given index%%assistant [index]:add- Concatenate content to existing assistant message
Example usage:
%%user 3:set
How do I read a CSV file?
%%assistant 3:set
Here's how to read a CSV file using pandas:
import pandas as pd
df = pd.read_csv('file.csv')
%%assistant 3:add
You can also specify additional parameters:
df = pd.read_csv('file.csv', encoding='utf-8')git clone https://github.com/yourusername/jupyter_whisper.git
cd jupyter_whisper
pip install -e ".[dev]"We welcome contributions! Please submit your Pull Requests.
MIT License - see LICENSE for details
Powered by:
- Claude by Anthropic
- OpenAI Whisper
- Perplexity AI
Made with ❤️ by Maxime
Note: This project is independent and not affiliated with Anthropic, OpenAI, or Perplexity AI.