🎤 Open-Source Voice Agent
100% open-source models · streaming overlap pipeline · English
STT
Whisper base.en
LLM
SmolLM2-1.7B
TTS
MMS-TTS-eng
VAD
Silero
🎤
START
Ready — click Start to begin
Conversation
Clear
Speech-to-Text
Whisper base.en
74 M params · GPU/CPU
Language Model
SmolLM2-1.7B-Instruct
1.7 B params · GPU/CPU
Text-to-Speech
MMS-TTS-eng
VITS · 16 kHz · CPU