We are building an Urdu-first voice AI system called Echopath, which relies heavily on accurate speech-to-text (STT) and text-to-speech (TTS) capabilities for effective voice agent interactions.

Please consider adding support for Groq’s Whisper Large V3 (and Whisper Large V3 Turbo) speech-to-text models as an additional transcription provider within the Vapi platform.

Rationale & Use Case:

Groq’s Whisper implementation offers state-of-the-art transcription accuracy with exceptionally low latency, made possible by Groq’s specialized LPU hardware. These models outperform many existing solutions in speed, accuracy, and multilingual capabilities, making them highly suitable for real-time and large-scale transcription use cases.

Groq's infrastructure enables speeds up to 216× real-time with consistent results across dozens of languages, making it an ideal addition for developers working with international voice agents and applications requiring fast turnaround and precise transcription.

Benefits to Vapi Users:

Ultra-fast transcription speeds suitable for real-time applications.

State-of-the-art accuracy, especially for non-English and accented speech.

Competitive cost-per-minute performance.

Strong multilingual and long-form audio support.

Feature Area: Transcription Providers / Speech-to-Text (STT) Integration