LiteRT.js not loaded

Copy Studio for LiteRT.js

Generate short copy from a topic or polish user-provided text. The AI flow is designed to run entirely in-browser with LiteRT.js once a compatible GenAI model is supplied.

Model Setup

GenAI wasm base path

Host the MediaPipe GenAI wasm folder locally for offline usage.

Model (.task/.litertlm) URL

Use MediaPipe GenAI LLM models. For best web results, prefer `*-web.task` (or `*-Web.litertlm`) files. WebGPU is required. The model file must be same-origin or hosted with CORS enabled.

Mode

Topic

Tone

Length

Output language

Max tokens

Output

AI output will appear here after the model runs.

LiteRT.js runs entirely on-device. No data leaves the browser.