Comfy-Org/ComfyUI · error · Exception

Seed Audio returned no audio (code={response.code}): {respon

Error message

Seed Audio returned no audio (code={response.code}): {response.message}

What it means

Raised after the Seed Audio API call succeeds at the transport level but the response object contains an empty `audio` field. The message includes the service's numeric response code and message, so the upstream refusal reason (content moderation, invalid reference audio, quota, etc.) is surfaced directly. This is an API-side failure, not a local validation error.

Source

Thrown at comfy_api_nodes/nodes_bytedance.py:3395

        response = await sync_op(
            cls,
            ApiEndpoint(path="/proxy/byteplus/api/v3/tts/create", method="POST"),
            response_model=SeedAudioResponse,
            data=SeedAudioRequest(
                model=model,
                text_prompt=text_prompt,
                references=references,
                audio_config=SeedAudioConfig(
                    sample_rate=int(sample_rate),
                    speech_rate=speech_rate,
                    loudness_rate=loudness_rate,
                    pitch_rate=pitch_rate,
                ),
            ),
        )
        if not response.audio:
            raise Exception(
                f"Seed Audio returned no audio (code={response.code}): {response.message}"
            )
        return IO.NodeOutput(audio_bytes_to_audio_input(base64.b64decode(response.audio)))


class ByteDanceExtension(ComfyExtension):
    @override
    async def get_node_list(self) -> list[type[IO.ComfyNode]]:
        return [
            ByteDanceImageNode,
            ByteDanceSeedreamNode,
            ByteDanceSeedreamNodeV2,
            ByteDanceSeedreamLayerSeparationNode,
            ByteDanceTextToVideoNode,
            ByteDanceImageToVideoNode,
            ByteDanceFirstLastFrameNode,
            ByteDanceImageReferenceNode,
            ByteDance2TextToVideoNode,

View on GitHub (pinned to 1c6d8d45b3)

Solutions

Read the code/message in the error text — it is the upstream service's own reason and dictates the fix.
Rephrase the prompt or swap the reference audio if the code indicates content policy or invalid input.
Verify the ByteDance/Volcengine API credentials and that the account has Seed Audio access.
Retry once after a short wait — transient service failures can return an empty audio payload.

Defensive patterns

Strategy: try-catch

Try / catch

try:
    audio = await seed_audio_node.execute(...)
except Exception as e:
    if "returned no audio" in str(e):
        # inspect code/message embedded in the error text, then retry once or surface to user
        log.warning(f"Seed Audio empty result: {e}")
        raise

Prevention

Keep prompts within the service's content policy to avoid non-zero response codes.
Use clean, audible reference audio clips (valid WAV/MP3, non-silent).
Confirm API credentials and Seed Audio entitlements before batch runs.

When it happens

Trigger: A sync_op call to the ByteDance Seed Audio endpoint returns 2xx with response.audio empty and a non-zero response.code — e.g. prompt rejected by content filters, unreadable reference audio, or service-side generation failure.

Common situations: Prompts with disallowed content; corrupted or silent reference audio clips; invalid API credentials hitting a service-level error path; transient upstream incidents; region/account restrictions on the Seed Audio model.

Related errors

AI-assisted analysis of Comfy-Org/ComfyUI@1c6d8d45b3 (2026-08-14). Data as JSON: /api/errors/c8d857b8be3ed5f0. Report an issue: GitHub.