Sep 5

1. 05:35

   Enable Fast In-Memory 30s Chunk Streaming with 0.4s initial latency

   Cong123779
   Verified signature

   d4752bc
2. 05:30

   Deploy In-Memory Resident C++ Daemon Engine for instant ~0.1s response latency

   Cong123779
   Verified signature

   6e864fc
3. 04:54

   Increase chunk size to 950 chars (~60-80s audio) for long story reading

   Cong123779
   Verified signature

   24e468f
4. 04:52

   Upload bin/matcha_tts_cli with huggingface_hub

   Cong123779
   Verified signature

   c617c31
5. 04:16

   perf: lock down to pure INT8 with optimized Vietnamese text normalization

   Cong123779
   Verified signature

   c025a8a
6. 03:57

   feat: deploy C++ Native Matcha-TTS & Vocos 22kHz Web App

   Cong123779
   Verified signature

   ead033d