Photograph a page you cannot read. Get it back in your language.
Last updated: 12 August 2026
Calque translates photographed documents and spoken phrases. This page says exactly what leaves your phone, who receives it, and what is kept. It is written to be read, not to protect us.
Zakaria Elkassmi, Oslo, Norway. Contact: amazigh.incognito@gmail.com
Your documents, their translations, your corrections and your page credits are stored on the device you used. There is no account, no sign-in, and no analytics or tracking of any kind. We cannot see your library, and we could not hand it over if we were asked.
On the device they are encrypted whenever the phone is locked, not only while it is switched off.
One thing worth knowing: if you have iPhone backup switched on, your documents are part of that backup, the same as your photos and messages. That copy is held by Apple or your computer under their terms, not ours, and we never see it. Turning off backup for Calque in your iPhone settings keeps everything on the phone alone.
Deleting a document deletes its images, its text and everything that came with it. Deleting the app deletes all of it, including the page credits you bought.
Documents. To translate a page, a picture of that page is sent to Anthropic (Anthropic PBC, USA), where it is read and translated. This happens only after you have agreed to it, and the agreement is asked for again if what we send ever changes.
Spoken phrases. In the Voice and sign features, the recognised text is sent to Anthropic to be translated. If the natural voice reads a translation out loud, that one sentence is sent to a speech provider (OpenAI, USA) to be spoken. Turning off “Natural voice” in Settings keeps speech entirely on the device, using the voice built into iOS.
Speech recognition. Turning your speech into text uses Apple’s speech recognition. On many languages this happens on the device. On the rest, Apple processes the audio on their servers, under Apple’s own privacy policy.
Nothing else is ever sent. Not your library, not your corrections, not which documents you have, not who you are.
Anthropic and OpenAI process the content to produce the translation or the spoken sentence, and do not use it to train their models. We do not store any of it on our side. Our server passes the request through and keeps no copy.
Because these providers are in the United States, this is a transfer outside the EEA. It happens on the basis of your consent, which you give before the first translation and can withdraw at any time by turning off cloud translation in Settings under More settings.
People use Calque for letters from hospitals, public agencies and immigration authorities. Those pages can contain health information and other sensitive details. That is exactly why the app has no account, no analytics, and no copy of your library. Send only what you are willing to have translated, and remember that the picture of the page is what travels.
Calque is not aimed at children and collects nothing that could identify anyone.
Because we hold no personal data about you, there is nothing for us to export or erase on request. Everything is in your hands: delete a document, or delete the app.
For the data you send onward, Anthropic’s and OpenAI’s own privacy policies apply.
If what the app sends ever changes, the consent screen changes with it and you are asked again. This page is updated at the same time.