Skip to main content
Each guide below shows how to move from another speech-to-text API to Ink. Ink has two realtime endpoints:
  1. Realtime STT (Auto): your client sends audio, and Ink finalizes transcripts automatically.
  2. Realtime STT (Manual): your client sends audio and signals when to finalize transcripts.
Most speech-to-text APIs combine both behaviors. Cartesia separates them to improve model performance. Use Realtime STT (Auto) for most agents. Many use cases need Realtime STT (Manual) instead, for example:
  1. Push-to-talk apps
  2. Pipelines where you know speech is over and are waiting for the transcript

Guides