ComfyUI Node
Vibe Music Engine
A ComfyUI node in VibeMusicEngine with 24 inputs and 11 outputs.
Vibe Music Engine
- audio
- beats_audio
- Lyrics
- frame_based_text
- Frame_long
- frame_numbers
- text_lines
- start_frame_numbers
- end_frame_numbers
- num_of_elements
- durations
- Start_ms
- Duration_ms
◄choose_modelbase►
◄source_languageauto►
◄target_languagesource►
◄continuation_text<LOGICAL CONTINUATION OF PREVIOUS VIDEO SHOOT>►
◄fps25.0►
◄words_per_line3►
◄prev_context_words0►
◄next_context_words0►
◄context_bracketstrue►
◄max_shot_length_frames125►
◄min_shot_length_frames25►
◄filename_prefixsc►
◄files_in_edlmp4►
◄beat_detectionfalse►
◄beat_sensitivity0.5►
◄beat_brutality0.50►
◄save_edloverwrite►
◄save_srtoverwrite►
◄srt_moderaw►
◄frame_gap1.0►
◄buy_me_a_coffeefalse►
◄usage_counterUsage count: 0►
CategoryVibeMusicEngine
Inputs (24)
| Name | Type | Default | Description |
|---|---|---|---|
| audio | AUDIO | Main audio input for speech-to-text processing. Should contain clear vocal content for optimal transcription. | |
| choose_model | COMBO | base | Select Whisper model size: tiny: Fastest, lowest accuracy base: Good balance of speed/accuracy small: Better accuracy, slower medium: High accuracy, slower large/large-v2: Best accuracy, slowest *.en models: Optimized for English |
| source_language | COMBO | auto | Select the source language of the audio: auto: Automatically detect language en: English pl: Polish de: German etc. |
| target_language | COMBO | source | Select the target language for transcription: source: Keep detected/specified source language en: English pl: Polish de: German etc. |
| continuation_text | STRING | <LOGICAL CONTINUATION OF PREVIOUS VIDEO SHOOT> | Text to be used for logical continuation of scenes. This text will be inserted when a scene is split into multiple parts. |
| fps | FLOAT | 25.00.1–120 | Frames per second for video output. Common values: 24fps: Film 25fps: PAL 29.97fps: NTSC 30fps: Digital 60fps: High frame rate |
| words_per_line | INT | 31–20 | Number of words to group per subtitle line. Lower values create shorter, more readable lines but more frequent cuts. |
| prev_context_words | INT | 00–20 | Number of previous words to add as context before the main text line. Helps with context understanding and prompt generation. |
| next_context_words | INT | 00–20 | Number of next words to add as context after the main text line. Helps with context understanding and prompt generation. |
| context_brackets | BOOLEAN | true | If enabled, adds [PREV: ...] and [NEXT: ...] brackets around context words. If disabled, adds context words without brackets. |
| max_shot_length_frames | INT | 125 | Maximum length of a single shot in frames. Longer shots will be split. At 25fps, 125 frames = 5 seconds. |
| min_shot_length_frames | INT | 251–100 | Minimum length of a single shot in frames. Prevents too rapid cuts. At 25fps, 25 frames = 1 second. |
| filename_prefix | STRING | sc | Prefix for output files. Will be used for both EDL and SRT files. Example: 'sc' will create 'sc_00001.mp4', 'sc.edl', etc. |
| files_in_edl | STRING | mp4 | File extension for video files referenced in EDL. Common formats: mp4, mov, avi, mxf. Must match your video workflow. |
| beat_detection | BOOLEAN | false | Enable automatic beat detection for music-synchronized cuts. Uses advanced audio analysis to find rhythmic patterns. |
| beat_sensitivity | FLOAT | 0.50.1–1 | Sensitivity of beat detection: 0.1: Only strongest beats 0.5: Balanced detection 1.0: Detect subtle beats Higher values may create false positives |
| beat_brutality | FLOAT | 0.500–1 | How strictly to align cuts to beats: 0.0: Subtle alignment 0.5: Balanced 1.0: Exact beat alignment Higher values may create jarring cuts |
| save_edl | COMBO | overwrite | EDL save mode: overwrite: Replace existing file no: Don't save EDL iterate: Create numbered versions (sc_001.edl, sc_002.edl, etc.) |
| save_srt | COMBO | overwrite | SRT subtitle save mode: no: Don't save SRT overwrite: Replace existing file iterate: Create numbered versions |
| srt_mode | COMBO | raw | Subtitle format mode: raw: Include all system messages clean: Only text content word: Word-by-word timing (karaoke style) original: Raw Whisper output all: Save all formats (raw, clean, word, and original) |
| frame_gap | FLOAT | 1.00.5–5 | Gap between words in frames when using word mode. Affects readability and timing precision. |
| buy_me_a_coffee | BOOLEAN | false | Support the development! ☕ By enabling this option, you'll be directed to a page where you can show your appreciation for this tool. Your support helps maintain and improve these nodes! |
| usage_counter | STRING | Usage count: 0 | Tracks the total number of times this node has been used. This helps us understand how the tool is being utilized and guides future development. The counter updates automatically with each use. |
| beats_audioopt | AUDIO | Optional separate audio input for beat detection. Useful when vocal track and music track are separate. If not provided, main audio will be used for beat detection. |
Outputs (11)
| Name | Type | Description |
|---|---|---|
| Lyrics | STRING | — |
| frame_based_text | STRING | — |
| Frame_long | INT | — |
| frame_numbers | STRING | — |
| text_lines | STRING | — |
| start_frame_numbers | STRING | — |
| end_frame_numbers | STRING | — |
| num_of_elements | INT | — |
| durations | STRING | — |
| Start_ms | STRING | — |
| Duration_ms | STRING | — |