Privacy Policy

Last updated: October 21, 2026

This Privacy Policy explains how Captio AI (“Captio AI”, “we”, “us”, or “our”) collects, uses, and protects your personal data when you use the Captio AI apps for iPhone and Android and the captioai.app website (together, the “Service”). Captio AI is built as an accessibility tool for deaf and hard of hearing people, and privacy is central to how it works.

1. Who is responsible for your data

The data controller responsible for your personal data is Nejc Dovžan Kukič, a sole proprietor established in Slovenia. You can contact us about privacy at any time at nejcdovzank@gmail.com.

2. Our privacy principles

In plain terms, this is how Captio AI treats your data:

• Your audio is processed in real time and is not stored on our servers.
• Your transcripts and summaries are stored securely to provide the Service — including syncing them across your devices — and are never sold or used to train AI models.
• We never sell your data.
• Your content is never used to train AI models, and it is never shared for advertising.
• On iPhone, we measure our advertising with Meta and OpenAI only if you allow tracking when the app asks. If you don’t, nothing is sent to them. The Android app does no advertising measurement at all.

3. Data we collect

• Account information. When you create an account with Sign in with Apple (on iPhone) or Google, we process your name, your email address, and an account identifier. If you use Sign in with Apple, Apple may provide a private relay email instead of your real one.
• Audio you capture or upload. When you use live captions, translation, or upload a file for transcription, your audio is processed to produce text. See section 4 for exactly what happens to it.
• Transcripts and summaries.Text generated from your audio. These are stored on your device and, so they’re available when you sign in on another device or reinstall the app, are also stored securely in our cloud, linked to your account. They are never used to train AI models.
• Subscription information. If you start a free trial or subscribe, we and our subscription provider, RevenueCat, process your subscription status and purchase history (plan, trial, renewals). To connect purchases to your account, RevenueCat also receives your account identifier, your IP address and, on iPhone, a device identifier provided by Apple. Payment is handled by the Apple App Store on iPhone and by Google Play on Android — we never receive or store your card details.
• App usage and diagnostics. Which features and screens you use, the app version, your device model and operating system version (iOS or Android), your approximate location (city and country, derived from your IP address), and a random identifier created by the app, together with crash reports and error codes. We use this to keep the app reliable and to improve it. It is processed by PostHog on servers in the EU and never includes the content of your audio, transcripts, or summaries.
• Advertising measurement (iPhone only).In the iPhone app we use these tools to learn which of our ads lead to installs and subscriptions (the Android app uses none of them: it never reads your device’s advertising ID and sends nothing to Meta or OpenAI):
  • Apple’s SKAdNetwork, which tells an ad network anonymously, at campaign level, that an install came from one of its ads. No identifier leaves your device.
  • An Apple Ads attribution token, which tells us whether the install came from one of our Apple Ads campaigns.
  • Only if you allow trackingwhen the app asks: the Meta SDK in the app sends app events (for example completing onboarding, creating an account, or opening the subscription screen) together with your device’s advertising identifier (IDFA), and RevenueCat sends trial and purchase events to Meta with the same identifier and Meta’s anonymous app identifier. You can change your choice at any time in iOS Settings → Privacy & Security → Tracking.
  • Only if you allow tracking when the app asks, and only if you came to us from one of our ads in ChatGPT: when you start a free trial or a subscription, we send that event (its type, plan, price, and country) to OpenAI together with your IP address and the click identifier from the ad, so OpenAI can count it as a result of that ad.
• Feedback you send us. Bug reports, feature requests, problem reports (with the error code shown in the app), transcript ratings and comments, and the optional in-app feedback form. The form can include whether you have hearing difficulties. Every question is optional.
• Emails. If you have an account, we use your email address, name, app language, and time zone to send you a welcome email, one follow-up email a week later, and a confirmation if you delete your account. Every email has an unsubscribe link.
• Invites. If you use the invite program, we process your invite code, who redeemed it, and a random identifier stored on your device, so that each device can redeem only one invite.
• Website analytics.Our website uses privacy-friendly, cookieless analytics that count visits, page views, and clicks on download buttons in aggregate. It does not use cookies, does not track you across sites, and does not build a profile of you. When you tap one of our ads in ChatGPT, the link opens our website, which sends you straight on to the App Store. On the way it notes which ad you tapped, OpenAI’s click identifier, and a one-way hash of your IP address (never the address itself), so that a trial you start later can be matched to the ad if you allow tracking in the app. This record is deleted after 8 days. If you accept a shared-folder invite on our website by signing in with Apple or Google, the sign-in uses your browser’s temporary session storage only while you join; you are signed out of the website right after, and nothing is kept in your browser.

Reminder notifications are scheduled on your device by the app itself. On iPhone, if you allow notifications, the app also registers a push token with our server so we can tell you when a new version is out or when someone adds a recording to a shared folder you are in; these are delivered through Expo’s push service and Apple. On Android, every notification comes from the app itself and no push token is registered.

Photos never leave your device. When you upload a video on iPhone, the app extracts its sound on your device and only the audio is sent for transcription. On Android, the video file you choose is sent to our speech-recognition provider, which uses only its sound and deletes the file as soon as the transcript is ready (section 4).

4. How your audio is processed

This is the most important part of how Captio AI works, so we want to be precise:

• Before anything is sent, the app asks you. The first time you use live captions (including the demo during setup), translation, an upload, or an AI action, the app tells you what is sent to a third-party speech-recognition provider and a third-party AI provider (both named below) and asks for your permission. If you say no, nothing is sent.
• When you use live captions, translation, or upload audio for transcription, your audio is transmitted securely to Soniox, our third-party speech-recognition provider, which converts it to text. The audio is processed transiently and then deleted — an uploaded file is deleted from the provider as soon as its transcript is ready. It is not stored on our servers, and it is not used to train any models.
• When Captio AI generates a title, a summary, or another AI action (key points, detailed notes, to-dos, questions and answers, a follow-up message), the transcript text is sent to Google’s Gemini API, our third-party AI provider, to produce it. We use a paid service under terms where your content is not used to train models. That provider may retain the input briefly for security and abuse-monitoring under its own terms, after which it is deleted.
• These providers are Soniox, Inc. (speech recognition) and Google LLC (Gemini API, used for titles, summaries and the other AI actions). They act as our processors, are bound by their own privacy terms and by data-processing agreements, and process your data only to provide the Service — never to train their models.

5. Where your transcripts and summaries are stored

Transcripts and summaries are stored on your device and are also synced to our secure cloud storage so they’re available when you sign in on another device or reinstall the app. They are linked to your account and protected by access controls so that only you can access them. If you put a recording in a shared folder, everyone in that folder can see its transcript and summary; leaving the folder, or removing someone from yours, takes that person’s recordings out of it. They are never sold or used to train AI models. When you delete a transcript, it is removed from your device and from our cloud; when you delete your account — which you can do at any time from within the app — all of your transcripts and summaries are permanently deleted.

Share links.If you choose Link when exporting a recording, we store a copy of its captions or summary (with its title, date, length and speaker names) and create a page at captioai.app that anyone with the link can open. Links use a long random code, so they can’t be guessed, and the pages are hidden from search engines. Nothing is shared unless you tap Link. The page stops working, and the copy is deleted, when you delete that recording or your account.

6. Legal bases for processing (GDPR)

• Performance of a contract — to provide the Service you request, including your account, syncing, and your subscription.
• Consent — for microphone access; for sending your audio to Soniox and your transcript text to Google (Gemini API), which the app asks before anything is sent; for advertising measurement with Meta and OpenAI on iPhone (your choice when the app asks to track, which you can change at any time in iOS Settings); and for the optional question about hearing difficulties in the feedback form (explicit consent under Article 9(2)(a) GDPR, which you can withdraw by contacting us).
• Legitimate interests— to keep the Service secure and reliable and to improve it (app analytics, crash and error reports), to measure our advertising with Apple’s privacy-preserving tools (SKAdNetwork and Apple Ads attribution) and by noting which of our ChatGPT ads brought you to our website (kept for 8 days), and to send the welcome and follow-up emails (you can unsubscribe at any time).
• Legal obligations — where the law requires us to process data.

7. Who we share data with

We do not sell your data. We share data only with the providers needed to run the Service, each for the purpose described in this policy:

• Soniox, Inc. — speech recognition (section 4).
• Google LLC — Gemini API for titles, summaries and the other AI actions (section 4), Google Sign-In if you choose it, and Google Play purchases on Android.
• Apple — on iPhone: App Store purchases, Sign in with Apple, push notifications, and Apple’s advertising attribution (SKAdNetwork and Apple Ads).
• Supabase — our database, accounts, and sync.
• RevenueCat, Inc. — subscriptions and free trials.
• PostHog — app and website analytics, on servers in the EU.
• Resend — email delivery.
• Expo (650 Industries, Inc.) — delivery of push notifications on iPhone.
• Vercel — website hosting and cookieless website analytics.
• Meta Platforms Ireland Ltd. — advertising measurement on iPhone, only if you allow tracking. Meta processes this data under its own privacy policy.
• OpenAI — advertising measurement for our ads in ChatGPT, on iPhone, only if you allow tracking. OpenAI processes this data under its own privacy policy.

We may also disclose data if required by law or to protect our legal rights.

8. International transfers

Some of our providers may process data outside the European Economic Area. Where that happens, the transfer is protected by appropriate safeguards, such as the European Commission’s Standard Contractual Clauses or an adequacy decision.

9. How long we keep data

• Audio — not stored on our servers; the recording is kept on your device for playback until you delete it.
• Transcripts and summaries — kept on your device and in our cloud (linked to your account) for as long as you keep them; deleted when you delete them or delete your account.
• Share links — the shared copy of a recording’s captions or summary, kept until you delete that recording or your account.
• Account data — kept while your account is active; deleted when you delete your account.
• Subscription records — kept as long as needed to provide your subscription and to meet legal obligations.
• Usage and diagnostic data — pseudonymous, and kept only as long as needed to run and improve the app.
• Feedback — kept until it has been handled; deleted sooner on request.
• ChatGPT ad clicks — the ad you tapped and a one-way hash of your IP address, deleted after 8 days.
• Advertising data sent to Meta or OpenAI — kept by them under their own retention policies.

10. Your rights

Under the GDPR you have the right to access, correct, delete, restrict, or object to the processing of your personal data, the right to data portability, and the right to withdraw consent at any time. To exercise any of these, email nejcdovzank@gmail.com. You can also delete your account, and all data linked to it, at any time directly in the app or on captioai.app/delete-account; turn tracking off on iPhone in iOS Settings → Privacy & Security → Tracking; and unsubscribe from emails with the link in any email. You also have the right to lodge a complaint with your local supervisory authority — in Slovenia, the Information Commissioner (Informacijski pooblaščenec).

11. Security

We use reasonable technical and organizational measures to protect your data, including encryption of audio in transit. No method of transmission or storage is completely secure. We keep audio out of server storage entirely, and your transcripts are held in access-controlled cloud storage that only you can access — both core parts of how we reduce risk.

12. Children

Captio AI is an accessibility tool intended for a general audience. If you are under the age of digital consent in your country (16 in some parts of the EU; 15 in Slovenia), you may use Captio AI only with the consent and involvement of a parent or guardian. We do not knowingly collect personal data from children without such consent; if you believe a child has provided us data without it, contact us and we will delete it.

13. Cookies and tracking

Our website does not use tracking or advertising cookies. The analytics we use are cookieless, which is why you will not see a cookie-consent banner. The app does not use cookies. The iPhone app tracks you for advertising measurement only if you allow it when iOS asks, and declining changes nothing about how the app works. The Android app does not track you for advertising.

14. Changes to this policy

We may update this Privacy Policy from time to time. The date at the top shows when it was last changed, and we will notify you of material changes where required.

15. Contact

For any privacy question or request, contact nejcdovzank@gmail.com.