The app listens to a speaker, turns the speech into text, translates it into other languages, and can read the translation aloud. A speaker can open a live room; listeners join with a code or QR code and follow along in their own language.

Your voice. When you press record or start a live room, the app uses your microphone. The audio is sent over an encrypted connection to our server, which sends it to OpenAI to turn it into text. We process the audio as it arrives and do not save your microphone recordings. The app asks your permission before the microphone is used, and before anything you say is first sent to OpenAI.

Camera (scanning a room's QR code). When you tap Scan QR code, the app uses your camera only to read the room's code. Nothing is photographed, saved or sent to us; the camera turns off when the code is read or you close the scanner. On the web version, starting the scanner downloads a free QR code reader (jsQR) from the jsDelivr network.

Profile picture and name. Everyone who creates an account adds a name and a profile picture, so people can see who is speaking. The picture is made small and square, and its location data is removed before it is stored on our server. You can change it or your name at any time in Your account; deleting your account deletes them.

Live video (rooms with Audio + video). If a speaker chooses Audio + video, their camera video is sent live through our video server to listeners who choose Video. It has no sound and is not recorded or stored. Listeners can stay on Audio only.

Public lives, the Feed and reports. A speaker can make a live Public (only people 18 or older, after accepting the Public rules). Then anyone can find it in the Feed: the speaker's name and profile picture, the live's name, description, cover picture, language and how many people watch are public. After a public live ends, anyone can read its captions; the speaker can delete them at any time from their profile. The name, description and cover are checked automatically by OpenAI's safety check before they are shown. Signed-in people can report a public live; a report keeps the reporter's account, the reason and what was shown, so we can review it. Private lives (the default) are never listed.

Following, reminders and notifications. If you follow a speaker or ask to be reminded of a live, we keep that with your account. On iPhone and Android, the app then asks to send you notifications ("… is live", "Starts in 15 minutes"); if you allow it, our server keeps your phone's notification address for this app only, sends it through Apple or Google, and removes it when you sign out. "Add to my calendar" stays on your device and is not sent to us.

Comments and ♥ in public lives. Comments you write in a public live are shown with your name and profile picture to everyone watching, translated into each person's language by OpenAI, and checked automatically by OpenAI's safety check. They are kept with the live until you or the speaker delete them, or the live's captions are deleted. ♥ are only counted, never linked to you.

Guests on stage. In any live, the speaker may let signed-in people 18 or older join the stage. A guest's speech is turned into text and translated like the speaker's, with the guest's name, and kept with the room's text. The people on stage hear each other's real voice (and see each other's camera in rooms with video) live through our video server; listeners see the cameras and get the translation. Nothing is recorded.

Blocks. If you block someone, we keep that with your account so they can't comment in your lives or ask to join your stage, and you don't see their lives. You can unblock them at any time.

Text of what was said, and its translations. In a live room, the text and its translations are stored on our server so listeners receive them and the speaker can see the room's history. This includes the optional topic line and any song lyrics the speaker sends verse by verse. When a room is reopened with the same code, earlier text stays stored (listeners no longer see it). Live-room text is kept until the speaker deletes their account (or, for a public live, deletes its captions). For single translations outside a live room, the text is sent back to your device and is not kept on our server. When you ask for a summary, the captions are sent to OpenAI to write it; the summary comes back to your device and is not kept on our server.

Spoken translations (voices). When a natural voice reads a translation aloud, the audio is created by OpenAI and kept on our server for about 10 minutes, so listeners can play it, then deleted automatically. The phone voice option is made by your own device or browser and is not sent to us or to OpenAI.

Account. You need an account to create live rooms and to use Talk One-to-One and Audience. Listeners who join a room never need one. We store your name, email address and password. Your password is stored scrambled (hashed), so we can't read it. Listeners never need an account.

Listeners. When you join a live room, the server records the language you chose and a temporary guest label (for example "Guest 4") for that room. We don't ask listeners for their name or email. A listener who is signed in can press Save to my account: the room's text in their language is then kept in their account until they delete it (Saved Translations → Delete) or delete the account.

Protected words and speaker details. In Settings you can enter names, an organization, a city and other words that must not be translated. They are saved on your device and sent with your translation requests so names are kept correctly. For a live room, they are saved with that room on our server.

Information kept only on your device. Your saved translations, settings and the sign-in for your account are kept on your device (or in your browser on the web). They are not sent to us unless described above. You can delete saved translations in the app at any time.

Technical information. Like most online services, our server sees your device's IP address when the app connects. We use it only to keep the service working and to stop abuse (for example, limiting how many requests come from one address). Server logs are kept for a short time for troubleshooting.

The app has no advertising and no tracking across other apps or websites. We don't sell or rent your information, and we don't use it to build advertising profiles.

We share information only with the companies that run parts of the service for us:

We may also share information if the law requires it, or to protect the safety of people using the app.

The app is not made for children under 13, and we don't knowingly collect personal information from them. An account must be created by someone 13 or older. If you think a child has given us personal information, write to interpreter-support@stoneintellect.com and we will delete it.

All connections between the app and our server are encrypted (HTTPS). Passwords are stored hashed. No system is perfectly secure, but we work to protect your information.

Our server is in Germany, and OpenAI processes data in the United States and other countries. By using the app, your information may be processed outside the country where you live.

If we change this policy, we will update the effective date above and, for important changes, tell you in the app.

Antonio Nsuami, doing business as Stone Intellect, Maine, United States — interpreter-support@stoneintellect.com