← Directory
Voice
Groq Speech
Ultra-low latency inference platform offering Whisper-based STT and high-quality TTS optimized for real-time voice apps.
Overview
Groq leverages its proprietary LPUs to deliver Whisper speech-to-text and text-to-speech APIs with single-digit millisecond latency. It’s designed for real-time voice agents where speed and consistency matter more than proprietary voice cloning.
Best for
- latency-sensitive voice agents
- real-time transcription
- fast TTS prototyping
Trade-offs
vs Deepgram
Groq wins on raw inference speed and free tier accessibility; Deepgram offers richer audio intelligence features like diarization and sentiment analysis.
What XeroHack pre-wires
When you pick Groq Speech in the interview, the engine emits these files into your scaffold:
- stt-whisper
- tts-voice
- realtime-voice
Coupon
No active coupon for this tool right now. Sign in to be notified when one is available.
Sign in