DIGIT ElevenLabs Sound Effects
Describe a sound, get a sound — ElevenLabs SFX inside your workflow
- audio
You've generated the video, and now it needs a sound. Not music, not a voice - a sound: a whoosh, a mechanical click, rain on a car roof, an 8-bit coin. That's the gap the DIGIT ElevenLabs Sound Effects node fills. You type a description, it calls ElevenLabs' SFX model, and you get an audio output ready to drop onto your clip.
It's part of the pack's ElevenLabs family, and it's the node you reach for when the sound design is too specific for a stock library search and too small a job to record. The whole thing is one text prompt away from a finished foley pass.
How it works
Give it text - "a satisfying metallic click followed by a soft spring release" - and three numbers do the rest. duration (default 5s, 0.5–30) sets how long the sound runs. prompt_influence (default 0.3) is the interesting one: it controls how strictly the model follows your text versus how literally it interprets it. Low influence, more interpretive; crank it when the model keeps ignoring your description. The loop toggle asks for a seamlessly looping sound effect, which is exactly what you want for ambient beds or UI sound design that repeats.
The optional api_key is auto-detected from ELEVENLABS_API_KEY (or the pack's DIGIT_ELEVENLABS_API_KEY) if you leave it empty. output_format defaults to pcm_44100 for maximum compatibility inside ComfyUI, with mp3_44100_192 and opus_48000_192 as lighter options for export.
Output is a single audio tensor - the same AUDIO type the pack's other audio nodes produce, so it plugs straight into a video mux, a save node, or whatever audio handling your graph uses.
Installing it
Standard pack install:
cd ComfyUI/custom_nodes
git clone https://github.com/thedepartmentofexternalservices/comfyui-digit.git
cd comfyui-digit
pip install -r requirements.txt
Or ComfyUI Manager → search comfyui-digit → install → restart. Then export your key before starting ComfyUI:
export ELEVENLABS_API_KEY=your_key_here
The honest notes
Two things worth knowing. First, this is a hosted, per-call service - ElevenLabs SFX is billed per generation, so the "describe and regenerate until it's perfect" loop costs real credits. prompt_influence is your efficiency knob: nudging it beats burning five generations on the same prompt. Second, remember this is a foley/sound-design tool, not a music tool - it makes sounds, not songs. If you need a track, that's a different model entirely. But for the thousands of small sound effects that make a video feel finished, having one inside ComfyUI beats tabbing out to a website and downloading files.
Inputs (6)
| Name | Type | Default | Description |
|---|---|---|---|
| text | STRING | Describe the sound effect to generate. | |
| duration | FLOAT | 5.00.5–30 | — |
| prompt_influence | FLOAT | 0.300–1 | — |
| api_keyopt | STRING | ElevenLabs API key. Auto-detected from ELEVENLABS_API_KEY env var. | |
| loopopt | BOOLEAN | false | Create a smoothly looping sound effect. |
| output_formatopt | COMBO | pcm_44100 | 3 options: pcm_44100, mp3_44100_192, opus_48000_192 |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| audio | AUDIO | — |