A little intelligence.
A lot of possibility.
Small, focused models for the things your app does every day. Built to work right where your users are.
LOCAL INTELLIGENCE
On your device.
Inference stays with the input.
Small, on purpose.
One focused task. A little model.
Yours to build with.
Readable weights. Reproducible recipes.
Little models.
Useful instincts.
Six starting points for more thoughtful software. Explore a model. Try it locally. See what it’s made of.
Explore all modelsLilt
146.5 KBFind the language behind the words.
Sprig
31.3 KBGive scattered ideas a place to belong.
Tinge
12.4 KBGet a feel for the words.
Hush
341 BExplore the space between sound and silence.
Tock
32.2 MBGive spoken words a place to stay.
Glyph
12.4 KBFind the shape inside the scribble.
Research preview. These are compact baselines, not production models. Sizes are measured artifact bytes. Read the evaluation notes ↗
Less talking.
More trying.
Give Lilt a sentence.
See which language it hears in the words.
Input stays here.
A small addition.
A new capability.
Load a model. Pass an input. Get a result.
One focused package for Node.js or Kotlin.
Local packages · Not yet published to registries
import { createModel } from '@embermote/lilt';
const model = createModel();
const input = "Les enfants jouent près de la maison.";
const result = model.predict(input);
console.log(result.label);
// Scores are not calibrated confidence.