Faster Whisper TransWithAI ChickenRice

Use it when

• You need to batch-convert Japanese audio or video into Chinese SRT/VTT/LRC subtitles on a local Windows GPU without writing any code
• Generic Whisper models give poor Japanese speech transcription or Japanese-to-Chinese translation quality and you want a purpose-tuned model with voice-optimized VAD, or a low-VRAM/CPU/cloud fallback when local GPU memory is insufficient

What it solves

Not the fit when

• General multilingual transcription pipelines embedded in code (use the faster-whisper library directly)
• Real-time streaming speech recognition
• Non-Japanese source audio where the bundled models offer no advantage
• cost-per-call-too-high
• multi-model-routing

Install

Download the release ZIP matching your GPU from GitHub Releases (CUDA 11.8/12.2/12.8 for NVIDIA, gfx-suffixed builds for AMD RDNA1-4, or CPU build) and extract; the Translate and Transcribe packages ship with the main Whisper model included

Invoke

Drag audio or video files onto the bundled batch scripts: the translate GPU .bat for Japanese-to-Chinese subtitles or the transcribe GPU .bat for Japanese transcription; outputs SRT/VTT/LRC. Optional Modal cloud inference via modal_infer.exe after 'modal token new'

Alternatives

No reviewed alternatives recorded yet.