Timeline Studio Skill
- routes Chinese and mixed Chinese/English narration through the owned Hojo TTS Light 80M FP16 browser bundle
- preserves stable narrator-to-voice selection across the two authorized built-in voices
- synthesizes long narration as separate sentence/breath-group assets
- standardizes adjacent voiceover spacing to 0.4 seconds on desktop and H5
- keeps MeloTTS out of the Chinese narration path
Editor
- adds Hojo WebGPU/WASM browser inference with ModelScope-first and Hugging Face fallback delivery
- adds sharded model packaging and provider-independent verified caching
- refreshes same-voice Chinese preview samples and removes the previous Kokoro Chinese worker
Validation
- full
npm run checkpassed - Skill structure validation passed