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 { useTranslation } from "react-i18next";
import { useMemo } from "react";
import type { ThreadChatApi } from "../../chat/useThreadChat";
import { MessageItem } from "./MessageItem";
export function MessageList({ chat }: { chat: ThreadChatApi }) {
const { t } = useTranslation();
const lastAssistantIdx = useMemo(
() => chat.messages.map((m) => m.role).lastIndexOf("assistant"),
[chat.messages],
);
return (
<div data-testid="message-list">
{chat.messages.map((m, i) => (
<MessageItem
key={m.id ?? i}
message={m}
streaming={chat.busy}
isLastAssistant={i === lastAssistantIdx}
busy={chat.busy}
onRegenerate={() => void chat.regenerate()}
onEditResend={(next) => void chat.editResend(m.id, next)}
/>
))}
{chat.busy && chat.messages[chat.messages.length - 1]?.role !== "assistant" && (
<p className="px-3 text-xs text-muted" data-testid="waiting-first-chunk">
{t("chat.waitingFirstChunk")}
</p>
)}
</div>
);
}