Privacy Policy

1. Who we are

MyInterpreter is a socialDNA product provided by SocialDNA Systems LLC, a limited liability company organized in Virginia, USA ("we", "us"). We decide how and why the personal data described here is processed, so we are its controller.

You can reach us, including our data protection contact (in Brazil, our encarregado), at privacy@socialdna.app.

2. The short version

• Your voice stays on your phone. During a call, your phone listens, translates and sends only the translated text of each phrase to the other phone, which speaks it with an interpreter voice. Your original audio is never sent to the other person or to our servers.
• Your voice signature never leaves your phone. It is kept encrypted in the phone's Keychain. We never receive it.
• Listen mode records nothing. The sound around you is processed on your phone and discarded right away.
• A conversation record exists only if both people accept it at the start of the call. Either one can say no, and the call goes on without a record.
• We keep a basic call record (who, when, how long, languages, voices and status) so your call history works.
• Chat is stored on our servers with its translations, so messages reach the other person and all your devices.
• We do not sell your data, we do not show ads, and we do not use your content to train AI models — ours or anyone else's.

3. What we collect and why

3.1 Your account

Name, email address, an optional phone number and your password. The password is handled by Firebase Authentication (a Google service) and we never see it in readable form. We use these to create and protect your account, let people you know find you by email, and send you account emails (verification, security alerts, changes to these documents).

3.2 Profile and contacts

Your language, the interpreter voice you chose (female or male) and the speech rate are kept on your phone. At the start of each call they are sent to the other participant's phone so it can speak your translation the way you chose. The contacts you add are saved on your phone. When you add someone by email, we check in real time whether that email belongs to a MyInterpreter user and do not keep the search. We never read your phone's address book.

3.3 Calls

• Call record (always): the account IDs of the caller and of the person called, their names at the time of the call, when the call rang, was answered and ended, its duration, the languages and interpreter voices used, where the interpreter ran, the status (for example answered, missed or declined), why it ended and the record level in force.
• Connection setup: to connect the two phones directly, each phone sends a short technical description of how to reach it (a WebRTC "offer" or "answer", which contains network addresses). Our server passes it to the other phone and erases it from the call record when the call ends. To learn its public network address, your phone also asks a STUN server, which sees your IP address at that moment.
• During the call: your microphone audio is processed on your phone — echo cancellation, the voice signature check, speech recognition and translation. Only the translated text of each phrase travels to the other phone, directly and encrypted, without passing through our servers. Without a conversation record, nothing said in the call is stored by anyone.
• Ringing: a push notification token (Firebase Cloud Messaging) so your phone can ring and show new messages. It is used only for MyInterpreter notifications.

3.4 Conversation record (optional)

Only if both participants accept it at the start of the call (section 12). Depending on the level, for each phrase it keeps: who spoke and from which device, the original language and the translation language, the original text and its translation, the times and — at the Text + audio and Certified levels — the clip of the speaker's voice that was transcribed (already cleaned by the voice signature) and the interpreter audio the listener actually heard. It also keeps a fingerprint (SHA-256 hash) of every phrase and clip, the engines and versions used, a voice-signature similarity score (a number from 0 to 1), the signed acceptance of each participant and any doubts you open about a phrase.

At the Certified level it also keeps: a chain in which each phrase points to the fingerprint of the previous one; a digital signature (ECDSA) made by each phone's secure chip; our server's arrival stamp for each phrase (time, IP address and network operator); a seal at the end of the call; an RFC 3161 timestamp; and an anchor in the Bitcoin blockchain through OpenTimestamps. Only fingerprints are sent to the timestamp services — never content.

3.5 Chat

Messages you write and their translations (stored beside the original, which is never changed), voice messages (the audio, its transcription and translations), documents and photos of documents you share (the original file, never altered, and its translated version), reactions, read receipts, presence and who is typing. While you type, a translated preview is sent to the other person; it is erased within 60 seconds, is never stored as a message, and you can turn it off in Profile. Translations are made on the sender's phone; when a device cannot translate (for example a computer, or a language the phone does not have), our own servers translate the text — never a third-party AI provider. Chat is encrypted in transit and at rest, but it is not end-to-end encrypted, because our servers must store it to deliver and translate it.

3.6 Listen mode

In Listen & translate, your phone hears the sound around you (a speaker, a film), translates it and speaks the translation in your earphones. The audio is processed on your phone and discarded right away; nothing is recorded or sent to us. In a conference room by QR, the organizer's phone translates the speaker and sends the translated text of each phrase through our servers to the listeners' phones. For a room we receive the translated text, the languages offered and chosen, and a temporary identifier for each listener; the room's text is deleted when the room closes.

3.7 Voice signature

A biometric voiceprint created and kept only on your phone, with your explicit consent. See section 11.

3.8 Device, security and quality

• Device registration: a device identifier, model, operating system and app version, and the public key of a signing key created in the phone's secure chip (Secure Enclave on iPhone, Keystore on Android). The private key never leaves the chip; not even we can read it. The public key lets anyone check the signatures on your consent and on certified records.
• Security logs: IP address, time, account ID, the request made and errors, to keep the service running, prevent abuse and investigate incidents.
• Call quality numbers: for each call, numbers such as the delay per phrase, how many pieces of audio were cut as echo, voice-signature similarity scores and phrases discarded by the echo filter. No audio and no text. We use them to tune echo protection and fix problems.

3.9 Support and purchases

What you send when you ask for help (your message, email address and any screenshots). When paid plans launch, Apple or Google will tell us the status of your subscription (product, dates, renewal and a transaction ID). We never receive your card or bank details.

4. What never leaves your phone

5. Legal bases (EU, UK and Brazil)

If the EU or UK GDPR or Brazil's LGPD applies to you, these are the legal bases we rely on:

Where we rely on consent, you can withdraw it at any time; this does not affect what was done before. Where we rely on legitimate interests, you can object (section 14).

6. Where your data is stored

• On your phone: profile, contacts, voice signature (Keychain, this device only), signing key (secure chip), language packs and the queue of record phrases waiting to upload.
• In our cloud, in the United States: our database on MongoDB Atlas; audio and documents in a private Google Cloud Storage vault, reachable only through short-lived signed links given to participants; our servers on Google Cloud; sign-in and notifications on Firebase (Google).
• Only between the two phones: the translated phrases of a live call, over an encrypted direct connection.

Every call and record in our database lists its owners by account ID. For calls and records, the owners are the two participants, and no other user can read them.

7. How long we keep it

8. Who we share it with

• The other participants. That is the service: your name, language and interpreter voice choice, the translated phrases of the call, your messages and their translations and — if both accepted — the conversation record.
• Service providers that process data for us, under contracts that limit them to our instructions (table below).
• Apple and Google as app stores. They deliver the app, deliver notifications to your phone and, when paid plans launch, bill you under their own terms and privacy policies. The speech, translation and voice engines built into your phone are set to work on the device, so your call audio is not sent to Apple or Google.
• When the law requires it. We disclose data when required by a valid legal order (subpoena, court order, warrant) or to protect someone's life or safety. We tell you first unless the law forbids it or it is an emergency. We can only hand over what we hold: we never have your call audio without a record, your voice signature or your private key.
• If the business changes hands. If SocialDNA Systems LLC is merged, acquired or sells MyInterpreter, your data may pass to the new owner under this policy; we will tell you before that happens.

9. International transfers

We are based in the United States and our cloud is in the United States. If you use MyInterpreter from another country, your data is processed in the United States, where privacy laws may differ from yours. When the EU or UK GDPR or Swiss law applies, our contracts with providers include the European Commission's Standard Contractual Clauses (and the UK Addendum). When Brazil's LGPD applies, we use the mechanisms of its Article 33, including the standard contractual clauses approved by the ANPD. You can ask us for a copy of these safeguards.

10. Security

• Connections to our servers use TLS 1.2 or later; the direct connection between the two phones is encrypted with DTLS.
• Data in our cloud is encrypted at rest. Audio and documents open only through short-lived signed links given to participants.
• Your voice signature is in the iPhone Keychain, readable only by MyInterpreter on that phone. Signing keys live in the secure chip and cannot be exported.
• Certified records are chained and signed, so any later change would be detected.
• Staff access is limited to people who need it, protected by multi-factor authentication and logged. Staff do not read or listen to your messages or records, except when you ask us to (for support), when the law requires it, or to investigate a serious security or abuse report.
• If a breach affects your data, we will notify you and the authorities as the law requires.

No system is perfectly secure. Protect your phone with a passcode and keep the app updated.

11. Voice signature: biometric data notice and consent

This section is our public written policy for biometric data, including for the Illinois Biometric Information Privacy Act (BIPA), the Texas Capture or Use of Biometric Identifier Act (CUBI), Washington's biometric identifiers law (RCW 19.375) and My Health My Data Act, the Colorado Privacy Act, the GDPR (special category data) and the LGPD (sensitive personal data).

What it is

When you set it up, you read a short text aloud for 20 to 30 seconds. The app turns that recording into a voiceprint — a set of numbers that describes the pitch (fundamental frequency) and timbre of your voice. The recording is deleted from the phone as soon as the voiceprint is computed.

Why we use it

Only to recognize your voice on your own phone during calls, so that echo, the interpreter's voice, a TV or other people in the room are cut before anything is transcribed. It is not used to identify you to us or to anyone else, to sign you in, for marketing or for any other purpose.

Where it is kept

In your phone's Keychain, encrypted and marked "this device only": it is not synced to iCloud Keychain, not included in backups and not moved to another phone. (On Android: encrypted storage protected by the Android Keystore.)

We never receive it

The voiceprint is never uploaded. We do not collect, receive, possess, sell, lease, trade, disclose or otherwise profit from it. To decide whether a piece of sound is you, the app briefly computes the same kind of numbers for each short piece of audio the microphone hears and compares them with your voiceprint; those numbers are discarded immediately, are never stored and are not used to identify anyone. If you accept a conversation record, each phrase stores only a similarity score (a number from 0 to 1), which cannot recreate your voice or your voiceprint.

Consent

The voice signature is optional. Before recording, the app shows this notice on its own screen and asks you to agree by tapping I agree; that tap is your written consent, signed electronically. Calls work without it, with weaker protection against echo. You must be 18 or older, and you may only enroll your own voice.

Retention and destruction schedule

The voiceprint is permanently destroyed at the earliest of: (a) when you tap Delete my voice signature in Profile; (b) when you sign out or delete your account; (c) the first time MyInterpreter starts after 12 months without use. If you delete the app first, iOS may keep the encrypted item in the Keychain, where no other app can read it; MyInterpreter erases it the next time it is installed and opened.

Withdrawing consent

You can withdraw consent at any time by deleting the voice signature. Nothing else in your account changes.

12. Calls, recording and consent of all parties

• Live interpretation is not a recording. Without a conversation record, nothing said in the call is stored by MyInterpreter, by us or by the other phone.
• Everyone is told. When a call is answered, both participants are told, in their own languages, that an AI interpreter is translating the call.
• A record needs both people. If a record level is set, both participants see "This conversation will be recorded (level). Accept · No record" in their own language when the call starts. If either one chooses No record, the call continues without a record. When the two accept different levels, the lower level applies. Each acceptance is signed by the phone and kept in the record.
• Designed for all-party consent laws. Requiring everyone's consent meets the strictest rules, including the US states that require all parties to agree to a recording (such as California, Florida, Illinois, Maryland, Massachusetts, Montana, New Hampshire, Pennsylvania and Washington), as well as Brazil and the EU.
• Accepting is never a condition of the call.
• Your part. Tell people near you when a record with audio is on. Do not record calls with other apps or devices without everyone's consent. Listen mode records nothing, and you must still follow the rules of the venue.

13. How MyInterpreter uses AI

• Speech recognition, translation and the interpreter voice are produced by AI models. On iPhone they are the engines built into iOS, set to run on the device; on Android they will be Google's on-device engines. Language packs are downloaded once from Apple or Google.
• The voice the other person hears is a synthetic interpreter voice chosen by the speaker (female or male). It is not the speaker's voice and it is not a clone of anyone's voice.
• AI translation can be wrong, incomplete or late. It is not a certified or sworn translation. That is why the original is always kept beside the translation in chat and in records, so a qualified or sworn translator can check it.
• When our servers translate chat text or documents as a fallback, they run models on our own infrastructure. We never send your content to third-party AI providers and never use it to train models.
• MyInterpreter does not make decisions about you that have legal or similarly significant effects.
• If we later offer an optional higher-quality engine that runs on our servers, your call audio would go to our servers for that call only. We will ask for your consent before turning it on and will update this policy first.

14. Your rights and how to use them

Everyone

Wherever you live, you can: access and download your data; correct it; delete it or your whole account; withdraw any consent; delete your voice signature; turn off the typing preview; and change or turn off conversation records for future calls.

EU, EEA, UK and Switzerland

Under the GDPR you also have the rights to restriction of processing, to data portability and to object to processing based on legitimate interests, and the right to lodge a complaint with your local data protection authority (in the UK, the ICO).

Brazil (LGPD)

Under Article 18 of the LGPD you may ask for: confirmation that we process your data; access; correction; anonymization, blocking or deletion of unnecessary or excessive data; portability; deletion of data processed with your consent; information about the entities we share data with; information about the possibility of not consenting and its consequences; and revocation of consent. You may also petition the ANPD.

California and other US states

Under the CCPA as amended by the CPRA, California residents have the right to know what personal information we collect, use and disclose; to delete it; to correct it; to opt out of its sale or sharing; to limit the use of sensitive personal information; and not to be discriminated against for using these rights. You can use an authorized agent. We collect these categories: identifiers (name, email, phone, account and device IDs, IP address); customer records; commercial information (subscription status, once paid plans exist); internet or other electronic network activity (call records, security logs, quality numbers); audio and electronic information (voice messages, and recorded clips only when a record is accepted); and sensitive personal information (account log-in and the contents of messages you send through the app). We collect them for the purposes in section 3, disclose them only to the service providers in section 8, and do not sell or share them. We use sensitive personal information only to provide the service, but you can still ask us to limit it. We honor Global Privacy Control signals.

Residents of Virginia, Colorado, Connecticut, Texas, Oregon and other states with privacy laws have similar rights, including the right to appeal our decision: reply to our answer with the word "Appeal". If we deny the appeal, you can contact your state Attorney General.

How to make a request

1. In the app: Profile › Privacy & data (download my data, delete voice signature, delete account).
2. By email to privacy@socialdna.app, from the address on your account or saying which account it is about.
3. We confirm it is you before acting (from inside your signed-in account, or with a code sent to your account email).
4. We answer within the time the law gives — 15 days in Brazil for a full access report, one month in the EU and UK (extendable by two months for complex requests) and 45 days in California (extendable by 45 days). Requests are free.

Records shared with another person

A conversation record belongs to both participants. You can remove a record from your history at any time; the content is deleted when both participants have removed it or its period ends, whichever comes first. If you want a record that includes you deleted for both, write to us: we will decide under the law that applies, taking into account the other participant's rights and any legal hold.

15. Minimum age and children

• You must be at least 16 to use MyInterpreter.
• You must be 18 or older to set up a voice signature, to turn on or accept a Certified conversation record and to buy a subscription (or have a parent or guardian do it, where the app store allows).
• MyInterpreter is not directed to children. We do not knowingly collect data from children under 13, or under 16 where the law sets a higher age of consent. If we learn that an account belongs to someone below the minimum age, we delete it. Parents or guardians who believe this happened can write to us.

16. Changes to this policy

We will update this policy as MyInterpreter changes. For important changes we will tell you in the app and by email at least 30 days before they take effect and, where the law requires, ask for your consent again. The version is shown at the top of this page; earlier versions are available on request.

17. Contact

• SocialDNA Systems LLC · Virginia, USA
• Email: privacy@socialdna.app
• Help center: MyInterpreter help

If you are in the EU, the UK or Brazil and you are not satisfied with our answer, you can complain to your data protection authority (in Brazil, the ANPD).