LiveKit avatar providers with Node.js support vs a Sume Node job

LiveKit's avatar models page lists 16 providers, all with Python plugins, and eight with Node.js plugins: Anam, Beyond Presence, D-ID, LemonSlice, Protoface, Runway, Tavus and TruGen (read 2026-10-05). If your backend is Node and you want a prepared clip instead of a live participant, Sume's job API needs only fetch.

The support matrix

The other eight (Avatario, AvatarTalk, bitHuman, Keyframe, LiveAvatar, Simli, Spatius and Synthesia) are Python-only on that page. Check the page again before you commit, because the table changes as plugins ship.

A Node job instead of a plugin

A Sume lip-sync call is plain HTTPS. Submit with an Idempotency-Key, poll the status endpoint until a terminal state, then read the result. This needs Node 18 or later, run as an ES module for the top-level await, or wrap it in a function.

const API = "https://api.sume.com";
const headers = {
  Authorization: `Bearer ${process.env.SUME_API_KEY}`,
  "Content-Type": "application/json",
};

async function call(method, path, body, key) {
  const h = key ? { ...headers, "Idempotency-Key": key } : headers;
  const res = await fetch(API + path, {
    method, headers: h, body: body ? JSON.stringify(body) : undefined,
  });
  if (!res.ok) throw new Error(`${res.status} ${await res.text()}`);
  return res.json();
}

const sub = await call("POST", "/v1/minimax/h3-max/lip-sync", {
  image_url: process.env.IMAGE_URL,
  audio_url: process.env.AUDIO_URL,
  duration_seconds: 5.84,
  resolution: "768p",
}, "node-lipsync-001");

let delay = 3000;
for (;;) {
  const st = await call("GET", `/v1/jobs/${sub.job.id}/status`);
  const state = (st.job ?? st).status;
  if (["completed", "failed", "canceled"].includes(state)) break;
  await new Promise((r) => setTimeout(r, delay));
  delay = Math.min(delay * 2, 30000);
}
console.log(await call("GET", `/v1/jobs/${sub.job.id}/result`));

Which to pick

Use a LiveKit plugin when the person on the other end must talk back in real time. Use a Sume job when the answer can be prepared: a greeting, a product explainer, a status update. The audio must be hosted on Sume media, and for H3 Max lip sync it must run 5 to 14.8 seconds. See measuring join and playback latency for judging the live side.