Whispering Tiger
A local AI transcription and translation tool for VRChat and streaming overlays that works offline after model downloads and supports CUDA acceleration.
Alternative to:
Whispering Tiger creates live transcripts and translations from audio on your computer for VRChat conversations and streaming captions. It can also read text in game images, so spoken dialogue and visible text can feed into the same translation workflow.
- Live captions and translation: Send transcribed or translated audio to VRChat through OSC or browser overlays through WebSockets, including overlays for OBS, vMix and XSplit.
- Speech and language models: Choose speech recognition models such as OpenAI's Whisper or NVIDIA's NeMo Canary, and translate text with NLLB-200 or M2M-100; automatic language identification is also supported.
- In-game text recognition: Capture and translate text from game images with EasyOCR, or recognize handwriting with GOT-OCR 2.0 and Microsoft's Phi-4 Multimodal LLM.
- Spoken output and voice conversion: Read transcripts and translations aloud with engines including Kokoro TTS and Silero, use F5/E2-TTS or Zonos for voice cloning and streamed playback, or convert live audio, files and generated speech with the optional RVC plugin.
The software is free and open source under the MIT license. Processing runs entirely on your machine and doesn't require an internet connection once the AI models have downloaded. CUDA GPU acceleration is supported and recommended. A separate native Whispering Tiger UI manages settings and application updates.