How can we help?

Start your first transcript

1. Open Eavesdrop and tap New transcript from the transcript library, then tap Start.
2. Read the on-device transcription explanation. Tap Allow, then grant microphone access and, on iPhone or iPad, speech recognition access.
3. Use a language available to your device's on-device speech recogniser. Start speaking when the recorder is listening.
4. Stop the session and choose Save to keep the text, or Discard to remove the draft.

Eavesdrop saves transcript text, not audio files. Check important details in the transcript before relying on or sharing it.

Microphone or speech permission was denied

Open your phone's system Settings, find Eavesdrop, and allow microphone access. On iPhone or iPad, also check Eavesdrop under Privacy & Security › Speech Recognition. Return to the app and try again. Device restrictions or organisation policies may prevent a permission from being changed.

An offline speech model is unavailable

Eavesdrop requires on-device recognition. It stops if your device cannot recognise the selected language locally. It does not fall back to cloud transcription.

iPhone and iPad

Language availability is controlled by iOS and depends on the device, system version, and available speech models. Eavesdrop has no model-download control on iPhone or iPad. Check the Recognition language setting in Eavesdrop and try another language supported on your device. A language appearing in the list does not guarantee that its offline model is available.

Android

Recording requires Android 12 or newer and a compatible dedicated on-device recogniser. Older devices can still open saved transcripts.

On supported Android 13 or newer devices, open Eavesdrop Settings and check the Speech model panel. If Download model is offered, connect to the internet and tap it. The system may schedule the download instead of completing it immediately; use Check again after it finishes.

Android 12 and some recognisers cannot report model status through the app. If the panel cannot confirm availability, check your device's speech or voice-input settings, or try another language. The settings available vary by device and recogniser. Downloading a model does not upload your microphone audio.

A session was interrupted

Return to Eavesdrop and check the recorder or transcript library. Unfinished transcripts are saved locally as recovery drafts and reopen paused. Resume when the app offers that action, or save the text already captured. A sudden app termination can lose words since the last completed save; a storage error may prevent a draft from being saved.

Edit, share, export, or delete text

Open a saved transcript from the library. Use Edit to correct its text, Share to send text through the system share sheet, or the overflow menu to export a text file or delete the transcript. Exporting and sharing are actions you choose; copies saved in another app are managed there.

Transcripts are stored locally and excluded from operating system backups. If you remove Eavesdrop and its data, we cannot restore your transcripts. Export anything you want to keep first.

Usage stats and privacy requests

Turn off Settings › Privacy › Anonymous usage stats to stop new analytics events and discard the unsent queue. This switch does not control diagnostic reports. See the privacy policy for the data included in each service.

For privacy questions or a deletion request, email privacy@aflutter.dev. We do not hold a server copy of your transcript content.