zhenzhen-suno-music-lowprice

Suno is the one music model people keep going back to the cloud for, and this node is the pack's entire Suno front door in a single box: 31 operations - generation, lyrics, import, extend, cover, reference generation, remix, sound effects, voice/timbre, stems, export, analysis, editing - all switching off one operation dropdown. Comfly_suno_music_lowprice runs it through the domestic low-price shop, so it takes an api_config from the pack's Comfly_seedance2_low_price_settings node like the rest of the low-price family.

The README is worth a read before you use it: the author verified all 31 operations against the real API and shipped 31 example workflows, one per operation, precisely because a node this broad is easy to misuse. The core rhythm to understand: many operations chain off a previous result via task_id. Generate a song → its task_id feeds the cover/extend/remix nodes. That's the pattern; the node's many required fields are mostly "which task, which version, which slice."

The inputs that matter

• operation - the big switch. The node shows only the parameters relevant to the selected operation, which keeps it usable.
• version - v3.5 through v5.5. The changelog calls v5.5 "效果真的太棒了" (genuinely great); if you're starting fresh, that's the one.
• prompt / style / title / tags / vocal_gender - the generation-side inputs.
• instrumental - instrumental-only toggle.
• task_id / task_id_2 - the chain links. Feed a previous run's task_id (output from this same node) into the extend/cover/remix-style operations.
• audio_index, continue_at, start_s/end_s, duration_s, speed - slicing and timing for the operations that need them (continue at bar X, extend from second Y, etc.).
• audio1–audio4 + audio_url1–audio_url4 - reference audio for import/cover/remix-style operations. The README notes local uploads must be at least 6 seconds, and voice/timbre samples 10–240s.

Outputs: audio1 and audio2 (the two track tensors), video (Suno makes a visualizer video - real tensor), text (for lyrics/describe ops), plus primary_url, result_urls, primary_path, result_paths, task_id (feed it forward!), and response. The pack deliberately does not save API keys or past results into the shipped workflows.

Installing it

cd ComfyUI/custom_nodes
git clone https://github.com/T8mars/Comfyui-zhenzhen

or ComfyUI Manager (search "Comfyui-zhenzhen"), restart, wire a settings node with your api.seedance.nz token into api_config. No model downloads.

Common issues

• Two identical songs - the README covers this: open the shop's async-task page, find the task, and grab the actual MP3s. The node sometimes returns a cached pair.
• Cover/remix "not taking" - the reference clip likely came from a different account; the README warns uploaded clips can fail cross-account. Re-upload under the same token.
• Upload rejected - check the length rules: ≥6s for general uploads, 10–240s for voice/timbre samples.

It's a lot of node, and it's exactly what it claims: Suno without leaving ComfyUI. If you mostly want one job, the pack's older single-purpose nodes are simpler - but if you want the whole Suno surface in one box, this is it.