Notebook Studio

Notebook Agent

draft mode

Agent source

Agent speed

Fast keeps prompts short and uses fewer response tokens.

Default is Qwen2.5 Coder 0.5B Q8_0 with the Fast profile for quicker browser notebook-cell generation.

Model settings

Hugging Face GGUF repo

GGUF file

GPU layers

Response tokens

WebMCP endpoint

Generated code is inserted below the selected cell.