Feature: Add SenseVoice for local voice input

[LauraGPT]

Important

Correction: The earlier comparative claims below are withdrawn; they were not established by a matched benchmark. SenseVoiceSmall supports Mandarin Chinese, Cantonese, English, Japanese, and Korean. It can emit language, emotion, and audio-event tags, but speaker diarization requires a separate model or pipeline (for example CAM++) and is not a built-in SenseVoice result. Runtime, timestamps, punctuation, and performance depend on the selected model, interface, hardware, and audio. FunASR and SenseVoice repository source code is MIT; model weights follow each model card. Please evaluate the exact integration on this project's workload.

Hi! VisionClaw is impressive — real-time AI for smart glasses with voice + vision.

For the voice input/ASR component, SenseVoice could improve responsiveness:

Why SenseVoice for smart glasses?

• 5x faster than Whisper — critical for real-time wearable interaction
• 234M params — lightweight enough for edge/phone processing
• Non-autoregressive — constant-time decoding, instant results
• Emotion detection — understand user's mood from voice
• Audio event detection — context awareness (music, traffic, crowds)
• 50+ languages — auto-detection

Edge deployment options

• Sherpa-ONNX: Optimized ONNX inference, Android/iOS/ARM support
  • https://github.com/k2-fsa/sherpa-onnx (5K stars)
• Python: pip install funasr

Links

• SenseVoice: https://github.com/FunAudioLLM/SenseVoice (8.3K stars)
• FunASR: https://github.com/modelscope/FunASR (16.7K stars)