Sonilo Text to Music
Prompt to a full track in under a minute, billed by the second
- AUDIO
The partner-node family has a music generator, and it's Sonilo. SoniloTextToMusic turns a text prompt into a complete track - up to six minutes of it - with the generation happening on Sonilo's servers through Comfy's API proxy while your GPU sits there doing absolutely nothing. If you've ever wanted background music for a video cut without licensing a track or training a model, this is the node.
Like every partner node it needs no Sonilo account or API key: you log in with your Comfy account (Settings → User) and spend prepaid credits (Settings → Credits). What sets this one apart is how it's priced: per second of generated audio, and the node's price badge literally scales with your duration slider. Longer track, bigger bill, in plain view. It landed in core around April 2026, so it's one of the newer arrivals.
The inputs
- prompt - the description of the music, up to 1,000 characters. "Energetic synthwave with a driving bassline, 128 BPM" is the right shape. More specific direction beats vagueness.
- duration - 1 to 360 seconds (6 minutes max), default 30. Your bill scales with this, so start short.
- seed - exists, and is honestly ignored. The tooltip says it plainly: the Sonilo service ignores the seed and it's only there for graph consistency. Don't chase reproducibility across runs; it isn't coming.
Output and wiring
One AUDIO output into Preview Audio to hear it or Save Audio (Advanced) to keep it. Because generation streams back in chunks, the node shows live progress text while it works - you'll see the track's title and the audio chunks arriving, which is a nice touch of visibility you don't get from most API nodes.
The honest gotchas
- The bill is the duration. A 6-minute track costs six times what the default 30-second clip does. Generate short, listen, then commit to the full length.
- The seed is a lie, sort of. It's ignored by design. If you need variation, re-run or change the prompt; if you need the same track twice, don't count on it.
- It's music, not stems or lyrics. You get a composed piece, not a master project. For section-level edits you're in a DAW, not a node.
- It's a paid cloud call -
--disable-api-nodeswill hide the whole partner family if you ever want it gone.
The fastest text-to-music path in the graph, and the pricing is at least transparent about it. Prompt it, preview it, save it.
Inputs (3)
| Name | Type | Default | Description |
|---|---|---|---|
| prompt | STRING | Text prompt describing the music to generate. | |
| duration | INT | 301–360 | Target duration in seconds. Maximum: 6 minutes. |
| seed | INT | 00–18446744073709550000 | Seed for reproducibility. Currently ignored by the Sonilo service but kept for graph consistency. |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| AUDIO | AUDIO | — |