How many voices can I clone at once?

This workflow supports up to three reference voices. Connect one to three LoadAudio nodes to the ByteDanceSeedAudio reference inputs in the order you want them used.

What makes a good reference clip?

Use clean, dry recordings with minimal background noise and reverb. A few seconds of natural speech with consistent tone works best. Avoid music, overlapping speakers, or heavy processing.

Does the model copy the exact words from the reference audio?

No. The reference clips transfer voice identity and style (timbre/prosody). The actual content comes from your text prompt, so write exactly what you want spoken and describe ambience or SFX clearly.

Do I need an internet connection or API key?

Some ByteDanceSeedAudio setups use a remote API. If so, you’ll need network access and to configure credentials as instructed by the node’s documentation. Check the node’s settings and README in your ComfyUI environment.

Seed Audio 1.0: Text + Audio to Audio

This ComfyUI workflow turns text plus one to three reference voice clips into a new, fully synthesized audio scene. At its core is the ByteDanceSeedAudio node (using Seed Audio 1.0), which extracts speaker identity from your reference clips and conditions generation on your text prompt. You can use this to produce dialogue in cloned voices, add ambient sound cues, and even light background music—all in one pass.

Technically, the flow is simple and efficient: LoadAudio nodes import your reference clips, the ByteDanceSeedAudio node fuses their speaker characteristics with your prompt to render new audio, and SaveAudioAdvanced writes the result to disk. A MarkdownNote in the graph provides inline guidance. Key parameters live in ByteDanceSeedAudio—enter your prompt, plug in up to three reference audios in order, set duration/seed, and adjust any available generation controls. SaveAudioAdvanced lets you pick output path and format so you can quickly iterate and export consistently.

API

Use this workflow from code

Every Comfy workflow is a JSON graph. The payload below is this workflow, exactly as ComfyUI runs it — fetch it from the URL, keep it in version control, or load it in ComfyUI and run it node by node.

d54ddc21563b.json

Fetching workflow JSON…

Run it from TypeScript or Python with the Comfy SDK. The same code targets Comfy Cloud or a ComfyUI you host yourself — only the base URL changes.

// Install (beta)
npm i @comfyorg/sdk

// Run this workflow (TypeScript)
import { Comfy } from "@comfyorg/sdk";

const client = new Comfy({ apiKey: "comfyui-..." });
const wf = await client.workflows.fromFile("workflow_api.json");
const job = await client.run(wf);
await job.getOutputs("<output-node-id>")[0].toFile("output.png");

The SDK takes a workflow in API format: open this workflow in ComfyUI and use File → Export Workflow (API).

Comfy Cloud API access requires a plan with an API key.

SDK docs Get an API key

FAQ

Frequently Asked Questions