This project integrates the tldraw whiteboard into Jupyter notebooks, enabling interactive visual data exploration and AI-assisted content generation. It targets data scientists, researchers, and developers working within the Jupyter ecosystem who need a dynamic canvas for their workflows. The primary benefit is a seamless, in-notebook whiteboard experience that can be enhanced with AI capabilities.

How It Works

The library leverages the @tldraw/tldraw JavaScript library, packaging it for use within Jupyter environments. It utilizes anywidget as its backend, ensuring that all necessary JavaScript and CSS are bundled with the Python package, eliminating CDN dependencies and improving reliability. For AI features like MakeReal, it integrates with OpenAI's GPT-4 Vision API, requiring an API key for image-to-prompt generation.

Quick Start & Requirements

• Install: pip install tldraw
• Run: Launch JupyterLab (jupyter lab) and import TldrawWidget from tldraw.
• Prerequisites: Python 3.11+, OpenAI API key for MakeReal functionality.
• Setup: Minimal setup, primarily requiring an API key for advanced features.
• Docs: https://github.com/kolibril13/jupyter-tldraw

Highlighted Details

• Integrates the latest @tldraw/tldraw versions.
• Supports AI-powered image-to-prompt generation via MakeReal with OpenAI.
• Uses anywidget for robust JS/CSS bundling.
• Includes experimental features like TldrawSegmentation and SVG/PNG export.

Maintenance & Community

The project has a history of frequent updates, with recent versions focusing on integrating newer tldraw releases and OpenAI models (e.g., GPT-4o). The changelog indicates active development.

Licensing & Compatibility

The project appears to be distributed under a permissive license, though the README does not explicitly state the license type. Compatibility with commercial or closed-source projects is likely, but confirmation of the specific license is recommended.

Limitations & Caveats

The MakeReal functionality is dependent on an external OpenAI API key and incurs costs. Some features are marked as experimental, suggesting potential instability or future changes. The project previously used ipyreact but migrated to anywidget for improved reliability.