Offline transcription
How to transcribe audio without an internet connection
Offline transcription means the speech-recognition model runs on your phone instead of sending the recording to a server. That matters when you have no signal — and when the recording is sensitive.
1. Download the speech model once
True offline transcription still needs an initial model download. Do this on Wi-Fi before travelling, entering a secure location or working somewhere with poor connectivity.
2. Record or import the audio
You can record a new conversation or import an existing audio or video file. The important distinction is where transcription happens, not where the recording originally came from.
3. Run speech-to-text on the device
The phone processes the audio locally. No network request is required for transcription once the model is installed, so airplane mode is a simple way to verify an offline workflow.
4. Search, correct and export locally
Once the transcript exists, review speaker labels, search the text and export the result. If privacy is the reason you chose offline transcription, check whether optional sync features are enabled before sharing or backing up files.
