Saymila is built so that your voice never leaves your iPhone. Here is exactly what goes through our server, and what stays there.
Last updated: November 2, 2026
Saymila is published by Guillaume Le Coz. For any question about your data: guillaume@guillaumelecoz.com.
Speech recognition is done by iOS, on your iPhone. No audio recording is sent or kept, by us or by anyone else.
To be cleaned up, the transcribed text is sent to our server, which passes it to OpenAI (United States) and sends the result back. Our server never records it: not in a database, a file or a log.
We ask OpenAI not to store it. OpenAI does not use it to train its models, but may keep it for up to 30 days to detect abuse, under its own policy.
There are also anonymous statistics (words per day and per writing style, for all accounts together), tied to no account.
Your dictation history stays in the app, on your phone, for as long as you choose: 24 hours, 7 days (the default), 30 days, no limit, or not at all. It is never sent to our server.
The keyboard does not record what you type: full access is only used to pass text between the keyboard and the app.
Subscriptions are handled by Apple. We never see your payment details: Apple only tells us the plan you bought and when it ends.
No advertising, no analytics tools, no sale of data.
This data is only used to run the service you use: signing you in, counting your words, applying your plan.
It is kept as long as your account exists. A session unused for 180 days expires. Deleting your account (My account › Delete my account) immediately erases all of it from our server; only the anonymous statistics remain.
OpenAI, to clean up the text; Apple and Google, to sign you in; Apple, for payments; our hosting provider, which runs the server. No one else.
Under the GDPR, you can access, correct, delete or export your data, or object to its processing, by writing to guillaume@guillaumelecoz.com. You can also contact the French data protection authority, the CNIL (cnil.fr).