Feature: Add SenseVoice/FunASR as local STT engine — 5x faster, 234M params

[LauraGPT]

Hi! EchoKit is a great open-source voice agent platform.

For the ASR/STT component, SenseVoice could significantly reduce latency:

Why SenseVoice?

• Non-autoregressive — complete transcription in single forward pass
• 5x faster than Whisper — critical for real-time voice agents
• 234M params — lightweight
• OpenAI-compatible API: funasr-server --device cuda serves at /v1/audio/transcriptions
• 50+ languages, auto-detection

Quick integration

from funasr import AutoModel
model = AutoModel(model="iic/SenseVoiceSmall", vad_model="fsmn-vad")
result = model.generate(input=audio_chunk)

Links

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