Balance Transcription Accuracy and Latency

Choose the right transcription mode for your UponAI agent based on your accuracy and latency needs.

This guide only applies to cascading agents. If you are using speech-to-speech models, this feature does not apply.

Real-time transcription is a trade-off between latency and accuracy. Using interim results gives the lowest latency but with a higher chance of errors due to less context. Waiting for results with more context improves accuracy but adds delay after the user stops speaking.

Transcription Modes

Optimize for Speed

Uses the latest interim results with a low endpointing setting for downstream processing. Best latency, slightly less accurate on entities like numbers and dates.

Optimize for Accuracy

Uses results with a higher endpointing setting, waiting longer with more context to generate more accurate transcripts. Incurs ~200ms additional latency.

Which Mode Should You Use?

Benchmarking shows that both modes have similar Word Error Rate (WER). The main difference is in capturing entities like numbers, dates, and proper nouns.

Use case	Recommended mode
General conversation, low latency priority	Optimize for speed
Capturing numbers, dates, or specific entities	Optimize for accuracy