Feature Request: Integrate FunASR / SenseVoice as ASR engine

[LauraGPT]

Summary

Would love to see FunASR and SenseVoice supported as ASR engines in the TEN Framework for conversational voice AI agents.

Why FunASR / SenseVoice?

• Real-time streaming ASR — FunASR has native streaming support with punctuation restoration and inverse text normalization, ideal for conversational agents.
• Ultra-low latency — SenseVoice-Small processes 10 seconds of audio in ~70ms, enabling responsive voice interactions.
• 50+ languages — Multilingual support out of the box.
• Fully offline — No API calls needed, runs locally with Apache 2.0 license.
• Industrial-grade — Used in production by Alibaba Cloud and 100K+ installations.

Key Models

Model	Specialty	HuggingFace
Paraformer	High-accuracy non-autoregressive ASR	iic/speech_paraformer
SenseVoice	Multi-task (ASR + emotion + events)	FunAudioLLM/SenseVoiceSmall
Fun-ASR-Nano	End-to-end, HuggingFace native	FunAudioLLM/Fun-ASR-Nano

References

• GitHub: https://github.com/modelscope/FunASR (16K+ stars)
• SenseVoice: https://github.com/FunAudioLLM/SenseVoice (8K+ stars)
• Streaming demo: https://funasr.com

Happy to help with integration!