feat(tts): add Qwen3-TTS as third TTS backend#311

Open

basnijholt wants to merge 18 commits intomainfrom

feat/qwen-tts-backend

Commits on Jan 25, 2026

feat(tts): add Qwen3-TTS as third TTS backend
basnijholt
committed
Merge 3cfcfa347b0acd0224989fb0a67695204ac91d43 into cb4293c67a4d41b1fbec2b24b909371d39fcf2ea
basnijholt
authored
Update auto-generated docs
github-actions[bot]
committed
fix(tts): add qwen-tts to requires_extras and add tests
basnijholt
committed
fix(tts): address PR review feedback for Qwen backend
basnijholt
committed
fix(docker): use COPY --chmod to avoid duplicate layer in transcription-proxy (#313)
basnijholt
authored
fix(tts): remove unused QWEN_DEFAULT_SAMPLE_RATE constant
basnijholt
committed
docs(skill): emphasize --from flag for branch-based work (#312)
basnijholt
authored
fix(tts): address PR review feedback - use helpers and clean up deps
basnijholt
committed
fix(tts): preserve original _extras.json format and ordering
basnijholt
committed
Merge d44164e6324f637a32650f149e8abcba3edcebd0 into 7a8c37ed1d09bfeff3795b8758cc4e2df8f2724c
basnijholt
authored
Update auto-generated docs
github-actions[bot]
committed
fix(tts): correct Qwen model size to ~4GB
basnijholt
committed
fix(tts): move first-party imports to module level in cli.py
basnijholt
committed
fix(tts): move pcm_to_wav import to module level
basnijholt
committed
fix(tts): remove flash_attention_2 requirement (not always available)
basnijholt
committed
perf(tts): add torch.compile() optimization for Qwen backend
basnijholt
committed
perf(tts): add Flash Attention 2 support for Qwen backend
basnijholt
committed