Cloud Speech-to-Text (v2.10)

You can now use cloud-based speech-to-text providers for real-time transcription during sessions. This provides transcripts on mobile devices without draining battery or using device resources. You are required to provide your own API key for your cloud provider of choice.

What it does:

Real-time PCM audio streaming to cloud providers
Support for multiple STT providers (OpenAI, Deepgram)
Automatic fallback if connection issues occur
Transparent to users - works just like on-device transcription

Benefits for mobile:

Better battery life - offloads processing to cloud
Works on devices that don't support larger on-device models

How to enable:

Go to Settings → Speech Recognition
Choose a cloud STT provider
Sessions will automatically use cloud transcription

Privacy note:

When using cloud STT, audio is sent to the selected provider for processing. You can switch back to on-device transcription anytime in settings.