Using speech-to-text with the Grammarly Keyboard for iOS, you can speak into any text field on your iPhone or iPad and get a clean, polished transcript without typing a single word.
Before using speech-to-text, make sure you have Grammarly installed and set the Grammarly Keyboard for iOS as your active keyboard on your device. Learn how to set up the Grammarly Keyboard in this article: How to install Grammarly for iPhone
What is speech-to-text?
Speech-to-text lets you talk naturally into any text field when using the Grammarly Keyboard. When you stop recording, Grammarly gives you a clean transcript without filler words and self-corrections, and with punctuation added. Your tone and meaning stay the same. Nothing is rewritten unless you decide to change it yourself.
How to use speech-to-text
Dictating your text
To start dictating, follow these steps:
- Tap any text field in any app to open the Grammarly Keyboard.
- Tap the microphone icon in the keyboard toolbar.
- After tapping the microphone icon, you’ll briefly see a Grammarly-branded screen before recording starts. Speak naturally at your normal pace.
Note: You’ll need to grant Grammarly access to your microphone before you can use this feature.
When you finish speaking, your transcribed text will appear in the text field.
Additional note for iOS 26: After the screen transition, you’ll need to manually swipe back to your app before you can start speaking. On earlier iOS versions, this happened automatically.
Refining your dictation with AI
If you’d like to adjust your dictated text after it appears, you can use the generative AI assistance built into the Grammarly Keyboard. You can ask it to:
- Shorten or expand your message
- Adjust the tone (more professional, more friendly, more direct, etc.)
- Rewrite or restructure the content entirely
You can also make manual edits directly in the text field, exactly as you would with your regular iOS keyboard.
Questions and answers
You can use speech-to-text in any app that supports the Grammarly Keyboard and has a text field.
The text will appear after you stop speaking. Grammarly processes your audio and returns the full cleaned-up text once you’re done speaking — it’s not a live word-by-word transcription. There’s a brief processing delay, which is normal.
Make sure Grammarly has microphone permission turned on and that you have a stable internet connection. If no audio was detected or the connection was interrupted, the transcription may return nothing.
Speech-to-text only transcribes and lightly cleans up your speech by removing filler words and adding punctuation. If you want to rewrite or adjust the tone, you can use Grammarly’s generative AI assistance separately.
No. Speech-to-text requires an internet connection to process audio and return transcriptions.
No. Grammarly does not store dictation history or voice recordings, so past transcriptions can’t be recovered.
On iOS 26, Apple turned off the automatic return behavior. You’ll need to manually swipe back to your app before speaking.
Yes. You can enable Background Microphone in Grammarly’s Settings to start recording without the transition screen.
No. The microphone only activates when you tap the microphone button. When recording is active, iOS shows an orange indicator at the top of the screen.
Audio is sent directly from your device to OpenAI for processing and is not stored or cached on Grammarly’s servers. No user or device identifiers are attached to the audio, and Grammarly does not retain voice recordings or use them for model training.