Run 35B MoE models in ~2.5 GB of RAM. Weights stream from SSD; a trained router predicts loads a step ahead. Mac, iPhone, Android, Windows supported
Python
// App-layer sampling presets. Engine default 0.7 is left alone for parity; precise 0.2 is the product default.
export type GenPreset = "precise" | "balanced" | "creative";
const GEN_KEY = "edge0.gen";
export const GEN_TEMPERATURE: Record<GenPreset, number> = {
precise: 0.2,
balanced: 0.7,
creative: 1.0,
};
export function storedGen(): GenPreset {
const v = localStorage.getItem(GEN_KEY);
return v === "balanced" || v === "creative" ? v : "precise";
}
export function setGen(p: GenPreset): void {
localStorage.setItem(GEN_KEY, p);
}
export function currentTemperature(): number {
return GEN_TEMPERATURE[storedGen()];
}