← Back to products
Productlaunched
Vision-LLM agent for reMarkable tablet that watches handwriting and draws LLM responses
Smart Remarkable is a Vision-LLM agent for the reMarkable tablet. It monitors handwriting and, when triggered, sends a screenshot to an LLM, then draws or types the response back onto the screen. It includes a Select Mode that lets users lasso a region of handwriting and get an LLM answer drawn into a chosen box, with real pen strokes that can be moved and resized.
Visit project ↗59 engagement·1 source·Tue, Jul 7, 2026, 06:06 PM
Forked from awwaiid/ghostwriter, extended with Select Mode. Users can trigger the agent to capture the screen, send it to an LLM, and have the response rendered as pen strokes on the tablet. The Select Mode allows lassoing handwriting and receiving an answer in a user-defined box. The answer is drawn as real pen strokes, enabling native reMarkable manipulation. Tech stack: Vision-LLM agent, reMarkable tablet. Traction: 59 stars on GitHub.