Dictee
Support
Dictee is a private, on-device dictation app for iPhone. Everything runs locally — your voice and transcripts never leave the device.
Getting started
- Open Dictee.
- Wait a few seconds while the speech model loads. This happens once per launch and is faster on subsequent opens.
- Tap the microphone button and start speaking.
- Tap Stop Recording when you are done.
- Tap Rewrite to polish the transcript in a style you choose: Corrected, Formal, Concise, or Friendly.
- Tap Play to hear the result read back, or copy and share it anywhere.
Requirements
- iPhone 14 Pro or later (6 GB of memory or more)
- iOS 18.6 or later
- Approximately 2 GB of free storage for the bundled AI models
The AI models run entirely on your device and need at least 6 GB of memory. On devices with less memory, Dictee shows a "Device Not Supported" message instead of running slowly. Older iPhones and the iOS Simulator are not supported.
Permissions
Dictee asks for microphone access so it can hear your voice. Audio is captured on your device only and is never uploaded to any server.
No other permissions are needed. Dictee does not access your contacts, photos, location, or any other data.
Privacy
Dictee does not collect any data. There are no accounts, no analytics, and no network calls related to your dictation content. All AI models are bundled inside the app and process your speech entirely on your device.
See our full Privacy Policy for details.
Troubleshooting
The app shows "Preparing models" for a long time
The first launch can take 10–30 seconds while the speech model loads into memory. Subsequent launches are faster because the model stays cached. Keep the app open and wait — a status message at the bottom of the screen will confirm when it is ready.
Dictation produces garbled or empty results
For best results: speak in a reasonably quiet environment, hold the phone within 30–50 cm of your mouth, and speak clearly without trailing off. Dictee currently supports English and German. Other languages may partially work but are not officially supported.
Rewriting takes a long time
The on-device language model is large. Rewriting a long passage can take 10–20 seconds depending on the device and the length of the transcript. Performance improves if you close other apps to free up memory.
The app crashes on launch
Make sure you have at least a few hundred megabytes of free storage available and that you are running iPhone 14 Pro or later on iOS 18.6 or later. If the problem persists, try restarting the iPhone and relaunching Dictee.
Text-to-speech sounds robotic or cuts off
The bundled TTS model is a compact on-device model — voice quality is functional rather than studio-grade. For very long passages, split into shorter segments for better playback.
Contact
Found a bug or have a question? Email us and we will try to respond within five business days.
info@speecon.com