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
import type { ThreadRow } from "../chat/ipc";
type Listener = () => void;
class ThreadStore {
threads: ThreadRow[] = [];
activeId: string | null = null;
private listeners = new Set<Listener>();
private loaded = false;
subscribe(fn: Listener): () => void {
this.listeners.add(fn);
return () => this.listeners.delete(fn);
}
setActive(id: string | null): void {
this.activeId = id;
this.emit();
}
async refresh(): Promise<void> {
const { threadList } = await import("../chat/ipc");
try {
this.threads = await threadList();
this.loaded = true;
} finally {
this.emit();
}
}
get isLoaded(): boolean {
return this.loaded;
}
private emit(): void {
for (const l of this.listeners) l();
}
}
export const threadStore = new ThreadStore();