9Graph nodes
8Node classes
12Inputs
2Models
Tags: 2026-sotaaudiocuefoleysfxsound-designstable-audio-3
Inputs (12)#
The typed parameter surface callers bind when they request this workflow. Enum options and numeric bounds are the values the workflow document declares.
promptstringrequiredDescribe the target sound effect
sfx_typeenumdefault foleyPrimary SFX family
options:
foleyimpactwhooshambienceinterfacecreatureduration_secondsintegerdefault 6min 1max 20Target duration in seconds
intensityenumdefault moderateSFX intensity envelope
options:
subtlemoderatestrongextremespatial_profileenumdefault roomTarget space and reverb feel
options:
dryroomhallcinematicaudio_prompt_hintstringdefault Optional prompt-side conditioning hint for texture/transient shaping
reference_audio_urlstringdefault Reference audio URL for style/texture conditioning where backend supports it
reference_audio_urlsarraydefault []Optional additional reference audio URLs for multi-reference conditioning
conditioning_strengthfloatdefault 0.55min 0max 2Strength of reference-audio conditioning
conditioning_modeenumdefault autoConditioning mode preference for backend-specific audio guidance
options:
autoreference-audiomelodystyle-transfervoice-cloneprovider_capability_modeenumdefault autoHow strictly provider support for reference-audio conditioning is enforced
options:
autorequireddisabledseedintegerdefault -1Random seed
ComfyUI node graph (9)#
The executable ComfyUI prompt graph: 9 nodes across 8 distinct node classes, wired by 10 data dependencies. Nodes tinted green come from a custom node pack this workflow declares; the rest are ComfyUI core / baked-community classes.
flowchart TD
n0["1: CheckpointLoaderSimple"]
n1["2: CLIPLoader"]
n2["3: CLIPTextEncode"]
n3["4: CLIPTextEncode"]
n4["5: EmptyLatentAudio"]
n5["6: KSampler"]
n6["7: VAEDecodeAudio"]
n7["8: AudioLoudnessNormalize"]
n8["9: SaveAudioMP3"]
n1 -->|clip| n2
n1 -->|clip| n3
n0 -->|model| n5
n2 -->|positive| n5
n3 -->|negative| n5
n4 -->|latent_image| n5
n5 -->|samples| n6
n0 -->|vae| n6
n6 -->|audio| n7
n7 -->|audio| n8
classDef custom fill:#12b886,stroke:#0b6b4f,color:#04110c;
class n7 custom;
Nodes (9)#
1CheckpointLoaderSimplecoreckpt_name = stable_audio_3_small_sfx.safetensorsoutputs:
MODELCLIPVAE2CLIPLoadercoreclip_name = t5gemma_b_b_ul2.safetensorstype = stable_audiooutputs:
CLIP3CLIPTextEncodecoretext = {{constructed_prompt}} tmplclip = ◂ node 2 · out[0]outputs:
CONDITIONING4CLIPTextEncodecoretext = clip = ◂ node 2 · out[0]outputs:
CONDITIONING5EmptyLatentAudiocoreseconds = {{duration_seconds}} tmplbatch_size = 1outputs:
LATENT6KSamplercoremodel = ◂ node 1 · out[0]positive = ◂ node 3 · out[0]negative = ◂ node 4 · out[0]latent_image = ◂ node 5 · out[0]seed = {{seed}} tmplsteps = 8cfg = 1.0sampler_name = lcmscheduler = simpledenoise = 1.0outputs:
LATENT7VAEDecodeAudiocoresamples = ◂ node 6 · out[0]vae = ◂ node 1 · out[2]outputs:
AUDIO8AudioLoudnessNormalizecustom packaudio = ◂ node 7 · out[0]target_lufs = -16.0outputs:
AUDIO9SaveAudioMP3coreaudio = ◂ node 8 · out[0]filename_prefix = sfx_sound_designerquality = V0Prompt construction#
template{prompt}, {sfx_type} sound effect, {intensity} intensity, {spatial_profile} space, production quality, clean transientsVariables (4)#
prompt{{prompt}} tmplsfx_type{{sfx_type}} tmplintensity{{intensity}} tmplspatial_profile{{spatial_profile}} tmplModels & dependencies#
Custom node packs (1)#
The non-core ComfyUI node classes this graph requires; the RunPod worker image the workflow runs on must bake or install a pack that provides every one of them.
AudioLoudnessNormalizeModels required (2)#
stable_audio_3_small_sfx.safetensorst5gemma_b_b_ul2.safetensorsOutput contract#
What a successful run of this workflow returns.
typeaudioformatmp3codecpcm_s16lechannels2descriptionMastered SFX waveformaudio_package{"mixdown": {"artifact_id": "audio_mixdown", "path_template": "audio/{{job_id}}/mixdown/sfx-master.wav", "format": "wav", "sample_rate_hz": 48000, "bit_depth": 24, "channels": 2, "codec": "pcm_s24le"}, "stems": {"format": "wav", "sample_rate_hz": 48000, "bit_depth": 24, "channels": 2, "artifacts": [{"stem": "fx", "artifact_id": "stem_fx", "path_template": "audio/{{job_id}}/stems/fx.wav", "required": true}, {"stem": "ambience", "artifact_id": "stem_ambience", "path_template": "audio/{{job_id}}/stems/ambience.wav", "required": false}], "archive": {"artifact_id": "stems_archive", "path_template": "audio/{{job_id}}/stems/stems.zip", "format": "zip"}}, "multi_track": {"format": "wav", "sample_rate_hz": 48000, "bit_depth": 24, "channels": 2, "artifacts": [{"track": "dialogue", "artifact_id": "multitrack_dialogue", "path_template": "audio/{{job_id}}/multitrack/dialogue.wav", "required": false}, {"track": "sfx", "artifact_id": "multitrack_sfx", "path_template": "audio/{{job_id}}/multitrack/sfx.wav", "required": true}, {"track": "music", "artifact_id": "multitrack_music", "path_template": "audio/{{job_id}}/multitrack/music.wav", "required": false}], "archive": {"artifact_id": "multitrack_archive", "path_template": "audio/{{job_id}}/multitrack/multitrack.zip", "format": "zip"}}, "loudness": {"artifact_id": "loudness_report", "path_template": "audio/{{job_id}}/analysis/loudness-report.json", "format": "json", "standard": "ebu-r128", "target_integrated_lufs": -16, "max_true_peak_dbtp": -1}, "compliance": {"artifact_id": "audio_compliance_report", "path_template": "audio/{{job_id}}/analysis/audio-compliance-report.json", "format": "json", "loudness_tolerance_lufs": 1.0, "true_peak_tolerance_dbtp": 0.3, "max_clipping_percent": 0.1, "enforce_format_normalization": true}, "cue_sheet": {"artifact_id": "cue_sheet", "path_template": "audio/{{job_id}}/metadata/cue-sheet.csv", "format": "csv", "include_timecode": true, "include_beat_markers": false, "include_sections": true}}Taxonomy & routing#
How the control plane classifies this workflow — from the committed workflow-taxonomy-registry.json. It drives the consistency / control surface the agentic director can exercise over the workflow.
assetFamilyfilm-vfx-element-libraryoutputPackageProfileaudio-master-profilecontrolModalitiescamera-lockcolor-script-lockdepth-constraintidentity-adapterlatent-reusemodel-lockpalette-lockpose-constraintprompt-template-lockreference-ensemblesampler-scheduler-lockscene-lockseed-lockstyle-anchorstyle-locktemporal-lock +1 moreconsistencyDimensionsmotioncolor-scriptnotesAuto-mapped from audio defaults