### Answer
Retell uses its own proprietary Automatic Speech Recognition (ASR) engine for speech-to-text conversion. [(1) ](
)
The ASR engine converts spoken language into written text in real time, forming the first step in enabling AI voice agents to understand and respond to callers.
Retell offers two transcription modes: [(2) ](
)
-
Optimize for speed (Fast & Accurate modes)
-
Optimize for accuracy (Accurate mode)
The ASR engine supports 50+ languages and has been optimized for telephonic conversations, which are challenging due to poor connectivity, background noise, and various dialects/accents.
You can also use
boosted keywords to bias the transcriber model toward domain-specific terms (names, brands, street names, etc.) for improved transcription accuracy. [(3) ](
)