Sonilo Music (BudgetPixel)
Sonilo Music — licensed instrumentals, exactly as long as you need
- audio
- video
If you've ever finished a video, opened the audio tab, and stared at a library of tracks that are all the wrong length or the wrong mood, Sonilo Music is the node that fixes that specific pain. It's BudgetPixel's text-to-music service: describe the track, get licensed instrumental or background music of an exact length - 5 to 360 seconds - as MP3 or WAV. For video soundtracks, ads, app sounds, or game music, that "exact length" guarantee is the whole point. No cutting a 3-minute song down to fit a 22-second spot; you just ask for 22 seconds.
What it costs and what you get
Billed per second of the requested duration at 4 credits per second with a 10-second minimum - so the cheapest job is 40 credits, and a 30-second default track is 120. That makes it one of the cheapest things in the BudgetPixel pack, which is handy when the rest of your video workflow is eating credits per second on the visuals. Output is MP3 by default; WAV if you want the lossless stem for mastering. The model is described as producing instrumental and background music, so don't come here expecting a vocal pop track - that's a different product entirely.
The inputs that matter
- prompt - required, up to 2000 characters. The tooltip's example is the right shape: "30 seconds of upbeat electronic instrumental for a product launch video, no vocals." Genre, mood, instrumentation, and what it's for all help.
- duration - 5–360 seconds; billed per second with the 10-second floor.
- format -
mp3orwav.
The outputs are the pack's audio oddity: the node returns two sockets - audio and video. For a text-to-music node the video output is just an unused placeholder (the extraction code returns None there); the one that matters is audio, which wires into ComfyUI's audio outputs or a player. You'll need a 2025+ ComfyUI for the core AUDIO type to exist.
How it works
Standard BudgetPixel mechanics: your prompt goes to the hosted API as a job, the node polls until it succeeds, and the audio downloads into your graph. There's no audio to upload here - text-to-music means the request is small and fast.
Install and key setup
ComfyUI Manager → BudgetPixel, or:
cd ComfyUI/custom_nodes
git clone https://github.com/budgetpixel-ai/comfyui-budgetpixel.git
pip install -r comfyui-budgetpixel/requirements.txt
Only requests as a dependency. API key from a paid plan via BUDGETPIXEL_API_KEY or a budgetpixel.json file in the pack folder; keys never land in workflow JSON.
Where people get burned
The 10-second minimum is the quiet billing rule - a 5-second sting still costs you 40 credits. Keep it in mind for short UI sounds (that's actually the SFX node's job anyway - Sonilo SFX has a 3-second minimum). And the shared pack gotchas apply: ComfyUI caches unchanged outputs, so force_rerun is what makes it actually generate again (and bill again), and the API needs a paid plan.
Inputs (4)
| Name | Type | Default | Description |
|---|---|---|---|
| prompt | STRING | Description of the music (up to 2000 characters; even a few words work), e.g. "30 seconds of upbeat electronic instrumental for a product launch video, no vocals". | |
| durationopt | INT | 305–360 | Length of the track in seconds (5-360). Billed per second with a 10-second minimum. |
| formatopt | COMBO | mp3 | Output audio format. |
| force_rerunopt | BOOLEAN | false | — |
Outputs (2)
| Name | Type | Description |
|---|---|---|
| audio | AUDIO | — |
| video | VIDEO | — |