Starstruk Logo MarkStarstruk

v0.2.23

PaddlePaddle/Paddlev0.2.23Jul 9, 2026by QuentinFuxa

AI Summary

This release adds Qwen3-ASR backends, introduces prebuilt Docker images, adds API token authentication, and fixes various bugs related to session stability, timeouts, and memory usage.

Key Highlights

  • New Qwen3-ASR backends (streaming and vLLM)
  • Prebuilt Docker images available
  • API token authentication support
  • Fixes for long session timeouts and memory growth

New Features

  • Qwen3-ASR backends (qwen3-streaming, qwen3-vllm, qwen3-vllm-metal)
  • Prebuilt Docker images on GHCR
  • Optional API token auth (`--api-token`)
  • Per-session language and translation target via WebSocket query params

Full Release Notes

Qwen3-ASR backends, latency work and bug fixes.

## Added
- Qwen3-ASR backends: `qwen3-streaming` (PyTorch, CUDA/MPS/CPU), `qwen3-vllm` (CUDA), `qwen3-vllm-metal` (Apple Silicon), via the new [qwen3-asr-causal](https://github.com/QuentinFuxa/Qwen3-ASR-causal) package. The causal streaming mode uses 3x less compute, at constant cost per audio second (English only).
- Prebuilt Docker images on GHCR: `ghcr.io/quentinfuxa/whisperlivekit`.
- Optional API token auth (`--api-token`).
- Per-session language and translation target via WebSocket query params.

## Fixed
- `mode=full` sessions no longer lose early lines after 5 minutes (#372).
- REST transcription of long files no longer times out silently (#374).
- qwen3-vllm word timestamps and CJK spacing (#375, thanks @samx81).
- Sortformer memory growth on long sessions (#349).
- Local Sortformer weights: `--sortformer-model-path` (#378, thanks @felixmr1).