On-device AI Models for JavaScript and TypeScript in the Browser and Node.js
npm packages that run models client-side in the browser and in Node, with TypeScript types included.
The JavaScript and TypeScript packages run in Node and the browser, scoped under @desert-ant-labs on npm. Most use transformers.js, ONNX Runtime Web, or LiteRT.js for inference.
Everything runs locally, so data stays on the device. Install with npm and import with full TypeScript types.
Available models
PII redaction in the browser via LiteRT.js, no server round-trip.
Multilingual emoji suggestions from text, 23 languages, in the browser or Node.
Content topics from any text across 101 languages, on device via WebAssembly and LiteRT.js.
Single-stroke shape recognition via a local WebAssembly pipeline with LiteRT.js.
Speech enhancement in the browser through ONNX Runtime Web.
Language identification for short text across 84 languages, fully client-side, no runtime.
Pricing
Every model is free up to 100k monthly active devices per SDK. Unlimited inference per user.