ComfyUI Node
ControlFoley Generate
A ComfyUI node in ControlFoley with 17 inputs and 4 outputs.
ControlFoley Generate
- controlfoley_model
- video
- video_input
- images
- audio
- sample_rate
- inference_time_sec
- peak_vram_gb
◄prompt►
◄negative_prompt►
◄duration10.0►
◄seed42►
◄num_inference_stepsfixed►
◄guidance_scale4.5►
◄mask_away_clipfalse►
◄cache_video_featurestrue►
◄staged_offloadtrue►
◄clip_batch_size_multiplier40►
◄sync_batch_size_multiplier40►
◄reference_audio_path►
◄image_fps24►
CategoryControlFoley
Inputs (17)
| Name | Type | Default | Description |
|---|---|---|---|
| controlfoley_model | CONTROLFOLEY_MODEL | — | |
| prompt | STRING | — | |
| negative_prompt | STRING | — | |
| duration | FLOAT | 10.00.7–30 | Text-only generation uses 10s by default. Video generation follows input length up to 30s. |
| seed | INT | 420–4294967295 | — |
| num_inference_steps | STRING | fixed | Use 'fixed' to keep the default step setting, or enter an integer from 1 to 100. |
| guidance_scale | FLOAT | 4.50–20 | — |
| mask_away_clip | BOOLEAN | false | — |
| cache_video_features | BOOLEAN | true | — |
| staged_offload | BOOLEAN | true | Move encoders to CPU during sampling when the ControlFoley source supports it. The public upstream source does not implement this; the option is then ignored and a console note is printed. |
| clip_batch_size_multiplier | STRING | 40 | Integer 1-80. Frames per CLIP encoder call = batch size * multiplier. Use 4-8 on low-VRAM GPUs. |
| sync_batch_size_multiplier | STRING | 40 | Integer 1-80. Frames per Synchformer encoder call = batch size * multiplier. Use 4-8 on low-VRAM GPUs. |
| reference_audio_path | STRING | — | |
| image_fps | FLOAT | 241–120 | Frame rate used when the optional IMAGE input is connected. |
| videoopt | CONTROLFOLEY_VIDEO | — | |
| video_inputopt | VIDEO | — | |
| imagesopt | IMAGE | — |
Outputs (4)
| Name | Type | Description |
|---|---|---|
| audio | AUDIO | — |
| sample_rate | INT | — |
| inference_time_sec | FLOAT | — |
| peak_vram_gb | FLOAT | — |