- Add the ready-to-use DramaBox LoRA training workflow and cover
- List the workflow in the README examples
- Keep DramaBox fine-tuning documentation general and capability-focused
- Add the DramaBox integration workflow and matching cover art
- Link the workflow from the README engine workflow table
- Highlight complete straight and curly quoted spans in the Multiline TTS Tag Editor
- Preserve existing tag colors and ignore unmatched or multiline quote pairs
Add the Fish Audio S2 Pro engine workflow with native and custom speaker modes, inline instruction handling, language prompting, compilation, quantization, and SRT support.
Document the workflow in the README and include a square cover image for the example workflow.
Add initial MOSS-TTS LoRA training support
Implementation details:
- add unified MOSS Delay 8B LoRA training flow and dataset preparation nodes
- add MOSS training docs and example workflow for clip-based dataset building
- add 4-bit MOSS training option for lower-VRAM LoRA training
- allow MOSS direct TTS inference without forced narrator fallback
Improve Text to SRT Builder TTS-ready segmentation and ASR timing reuse
Implementation details:
- Add language-aware subtitle heuristic profiles with Portuguese (Brazil) defaults
- Add TTS-ready subtitle mode and preset for single-line semantic cue building
- Improve dangling-tail and incomplete-thought merging in SRT builder
- Serialize ASR timing output as reusable JSON and rename builder input to asr_timing_data
- Remove the duplicate SRT Advanced Options node alias and update bundled workflows
Add Qwen3 + ASR example workflow and improve Qwen3-TTS README documentation
- Add Qwen3 integration + ASR example workflow to Example Workflows table
- Clarify Qwen3-TTS is a single engine with 4 model types (CustomVoice, VoiceDesign, Base, ASR)
- Document ICL vs X-Vector cloning modes with reference transcript requirement
- Add style instruction limitation warning (Base model ignores instruction field)
- Note VoiceDesign is 1.7B only
Add VibeVoice wildminder format compatibility
Technical changes:
- Implement [1] [2] [3] [4] numeric character tag support for native multi-speaker
- Replace synthetic voice samples with None for true zero-shot generation
- Increase CFG scale limit from 5.0 to 10.0
- Add direct Speaker mapping for numeric character tags
- Maintain full backward compatibility with existing character systems
Add missing F5-TTS workflow and improve documentation
Technical fixes:
- Add comprehensive tooltips to Phoneme Text Normalizer node inputs
- Add Text Processing Nodes section to PROJECT_INDEX.md
- Document 📝 Phoneme Text Normalizer with all 4 processing methods
- Add missing workflow file reference for user testing
Fix IndexTTS-2 extra_model_paths integration
Technical details:
- Update IndexTTS-2 downloader to use extra_model_paths configuration
- Implement local: prefix pattern matching F5TTS/ChatterBox style
- Add auto-detect model discovery across configured paths
- Fix model path resolution for both downloads and local models
Fix IndexTTS-2 device mismatch after unload and add workflow
Technical fixes:
- Add IndexTTS-2 recovery logic in model manager for device synchronization
- Clear cached model config and force device reload after CPU offloading
- Handle both normal reuse and force_reload scenarios
- Add IndexTTS-2 integration workflow to README examples
Fix VibeVoice Native Multi-Speaker voice reference handling
Technical improvements:
- Add voice reference validation and normalization in native mode
- Detect invalid character name connections with helpful error messages
- Add auto-conversion fallback for misconnected inputs
- Implement global character-to-speaker mapping for SRT consistency
- Remove redundant debug output to clean up console logs
- Add auto-detection for manual Speaker format with mode switching
Centralize numba compatibility system with smart testing
Technical implementation:
- Replace 15+ duplicate JIT workaround blocks with unified system
- Add utils/compatibility/numba_compat.py with startup testing
- Implement fallback for environments without compatibility module
- Test librosa.stft() at startup and apply workarounds only when needed
- Eliminate contradictory startup messages between init and nodes
- Maintain backward compatibility with legacy Python 3.13 fixes
Reorganized 19+ scattered categories into 6 clean, logical categories:
- TTS Audio Suite/🎤 Text to Speech: Main user-facing nodes (TTS Text, TTS SRT, Voice Changer)
- TTS Audio Suite/⚙️ Engines: Engine configuration nodes for ChatterBox, F5-TTS, Higgs Audio, RVC
- TTS Audio Suite/🎭 Voice & Character: Voice management and RVC character models
- TTS Audio Suite/🎵 Audio Processing: Audio tools, analyzers, and processing nodes
- TTS Audio Suite/👄 F5-TTS: F5-TTS specific editing nodes
- TTS Audio Suite/🎬 Video Analysis: Video mouth movement and viseme detection nodes
Improvements:
- Consistent naming hierarchy with proper emoji placement on category names
- Logical grouping by functionality instead of engine type
- Moved video nodes from generic "image/video" to TTS-specific context
- Eliminated inconsistent emoji usage and random category variations
- Better user experience with intuitive node organization
From messy scattered categories to clean professional structure.
- Update README with new TTS/ folder organization for all model types
- Add TTS/chatterbox/, TTS/F5-TTS/, TTS/RVC/, TTS/UVR/ as recommended locations
- Maintain backward compatibility with legacy model locations
- Add PROJECT_INDEX.md documentation for missing download components
- Include HuBERT downloader and universal model downloader in project index
- Update unified voice changer workflow
Remove duplicate pitch_detection from RVC Engine and improve UI consistency
Cleaned up RVC Engine node by removing duplicate pitch_detection parameter that belonged in RVC Pitch Options.
Added sliders to RVC Pitch Options for better user experience and improved tooltips.
- Removed pitch_detection from RVC Engine INPUT_TYPES and function signature
- Added default 'rmvpe' fallback when RVC Pitch Options node not connected
- Added sliders to crepe_hop_length, filter_radius, and pitch_guidance for visual feedback
- Enhanced pitch_detection tooltip with comprehensive method descriptions
- Added new 'Unified Voice Changer - RVC X ChatterBox' workflow to README
- Created new 'Unified Workflows' section in README for modern v4+ workflows
Improves separation of concerns and provides better user experience with clearer parameter organization.
Add F5-TTS SRT workflow and fix README workflow links
- Added new F5-TTS SRT and Normal Generation workflow
- Fixed broken SRT workflow link in README (missing emoji prefix)
- Updated workflow section to properly categorize Advanced workflows
- All workflow links now point to correct files
- Add speed control dropdown widget (0.25x to 2x) to playback controls
- Implement setPlaybackSpeed() method using HTML5 Audio playbackRate
- Position speed control between time display and selection display
- Provide user feedback when speed is changed
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Fix waveform data display with real audio (30s duration vs fake 10.79s)
- Implement working cache system using web directory for JS-Python data transfer
- Add green region overlays showing detected silence/speech segments
- Fix RMS line zoom/pan compatibility with different data structures
- Consolidate UI layout with single-row button arrangement
- Reduce container height to fit properly within ComfyUI nodes
- Clean up debug logging for better console output
- Implement robust error handling for visualization rendering
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Renamed nodes with '(diogod)' suffix to avoid conflicts with original ChatterBox Voice repository
- SRT Voice TTS node remains unchanged as it was unique to this implementation
diodiogod