Repository Issue Activity (beta)

#537 Chatterbox in native C++/GGML, 6.68x faster than real time

#517 Language request: Catalan (ca)

#536 BUG: Persistent low-level background tone/hiss in generated audio (~4900 Hz + elevated silence floor), traced to the vocoder, not the input

#271 Extra noise at the end of the generated clip

#535 Feature: chatterbox vllm real-streaming support

#534 Feature Request: Add CLI support for TTS, Turbo, Multilingual, and Voice Conversion

#533 Bug: CPU loading fails when device=torch.device("cpu")

#193 Ways to reduce latency

#531 Missing API - The model silently fails and produces audio with long silences, happens to be at 1000/1000

#530 `t3_model` is no longer an expected argument, should remove from README.

#529 Enterprise High quality dataset features

#527 mtl_tokenizer.json missing Persian characters: آ (U+0622), أ (U+0623), إ (U+0625)

#526 Czech Language Support for Wake Word Training

#508 Chatterbox randomly takes forever to generate TTS

#256 Polish language has issues with saying letters like ś,ń,ż,ą,ę , maybe its something about the fonts decoding ?

#525 output_attentions=True forces eager attention, blocking SDPA/flash attention for English models (fix unreleased on master)

#524 Feature: Consider FunASR/SenseVoice for training data annotation

#367 "pip install chatterbox-tts" - DOESN'T WORK due to pkuseg==0.0.25

#523 Bug: mask.py uses logging.warning without importing logging module

#522 Does this support Indonesian language?

#515 Problem initializing multilingual model

#181 Provide a wrapper method that can take long lines to text and work around the 1000 token / 40 second limit

#521 Benchmark with your previous version

#47 One click installer - PySide GUI - Should Support PC/MAC/LINUX

#519 long_tail detection triggers too early, causing token repetition before forced EOS