Privacy Policy

Last updated: 2026-10-19

Who is responsible for your data

OwlPrompter is published by TECHNINN SOLUTIONS BİLİŞİM TEKNOLOJİLERİ A.Ş., İnkılap Mah. Sokullu Cad. No:5 Ümraniye / İstanbul, Türkiye ("we", "us"). For the purposes of the EU/UK General Data Protection Regulation and the Turkish Personal Data Protection Law (KVKK), we are the data controller for the limited processing described below.

You can reach us at support@techinn.solutions.

The short version

• Your scripts, recordings, subtitles and Delivery Score reports are stored only on your device.
• We operate no account system and no cloud storage for your library. We cannot read your scripts or recordings, because we never receive them — unless you start one of the cloud AI features, which sends only what that feature needs.
• The cloud AI features — AI eye contact, AI captions, Create with AI and AI rewrite — are optional and never run on their own. Each one tells you what it will send and asks first; uploads are deleted within 24 hours and are never used to train AI models. One switch in Settings turns them all off.
• We use privacy-first, anonymous product analytics (no personal data, no IP stored), Apple’s privacy-preserving Search Ads attribution, and Meta’s SDK to measure our ads on Facebook and Instagram. We show no ads inside the app and never ask for App Tracking Transparency permission, so on iPhone the advertising identifier (IDFA) is not collected. The effects SDK (DeepAR) also reports anonymous usage — details below. Our website uses Google Analytics and, in Europe and Türkiye, asks before setting any cookie.
• Personal data reaches third parties on our behalf only to sell you a subscription or AI credits, to run the cloud AI features you start, and to measure our advertising (Meta, through its SDK and through RevenueCat). The support chat is the other exception, and only if you open it — details below.
• Speech recognition and live streaming involve other companies — Apple, Google and whichever platform you broadcast to. Those sections below explain what leaves your device and when.

What stays on your device

The following are written to your device's private app storage and are not transmitted to us, unless you start a cloud AI feature on one of them (see “Cloud AI features” below):

• The scripts you write or import from PDF, DOCX, TXT or RTF files.
• Video and audio you record in the app, together with any .srt subtitle file exported alongside them.
• Your Delivery Score reports — eye-contact percentage, pace consistency, pauses and filler-word counts. The analysis runs entirely on your device.
• Your reading preferences, fonts, filters and virtual-background choices.

Deleting the app removes all of it. Because we hold no copy, we also cannot restore anything for you.

Camera and microphone

Camera and microphone frames are processed on your device to render the preview, burn in filters, virtual backgrounds and subtitles, and to detect whether you are looking at the lens. Face detection and background segmentation use on-device models (Apple Vision on iPhone, Google ML Kit on Android, and on both the DeepAR SDK for background and face effects); no image ever leaves the device for that purpose.

Video is written to a file only when you press record, and sent over the network only when you start a live broadcast or a cloud AI feature that needs it.

Speech recognition

Voice-tracked scrolling and automatic subtitles use the speech recognition built into your operating system. We ask it to run on-device wherever that is possible: on iOS we set requiresOnDeviceRecognition when the selected language supports it, and on Android we request offline recognition.

On-device recognition is not available for every language, device or OS version. Where it is not, your operating system may send short audio segments to Apple or Google for transcription under their own privacy policies. We never receive that audio or the resulting text.

If you would rather no audio ever leave your device, use the speed or timer scrolling modes and turn subtitles off.

Cloud AI features

OwlPrompter AI adds features that run on servers we operate rather than on your phone: AI eye contact, AI captions, Create with AI (script drafts) and AI rewrite. A new user can also try eye contact once on a short test clip during setup, and anyone can make a few free script drafts. All of this is optional. Nothing is sent for these features until you start one, and the first time you use each of them the app shows what will be sent and asks for your permission. These features are available in the iPhone and iPad app.

What is sent, feature by feature:

• AI eye contact — the recording you choose, picture and sound, so your face and your voice, to make a copy in which your eyes look at the camera. In the setup test, only that test clip of up to 8 seconds.
• AI captions — only the sound of the recording you choose (your voice); the picture stays on your phone. If the take is linked to a script and you let the captions use it, the script text is sent too, so the captions follow its spelling. If you ask for translations, the transcribed text is translated.
• Create with AI — what you type: the topic, any notes or text you paste, the format you pick (platform, length, tone, language, audience, template) and the opening line you choose.
• AI rewrite — the text of the script you ask to rewrite and the kind of change you asked for.

What comes back — a corrected video, captions with word timings, or script text — is downloaded to your phone and stored there like everything else you make in the app.

To run these features without an account, the app signs in to Firebase anonymously, which creates a random user ID. If you buy OwlPrompter AI or AI credits, your RevenueCat purchase identifier is set to that same ID, so your credits follow your subscription. For the one-time setup test and the free script drafts, the app also keeps a random install ID in your device's keychain, so these free allowances can be used once per device even after a reinstall. When a job is running, the app sends its push token so it can tell you when the result is ready. Requests are checked with Firebase App Check, which uses Apple's App Attest to confirm they come from the genuine app. None of these identifiers contains your name, your email address or your advertising identifier.

You can withdraw the upload permissions, or turn cloud AI off completely, at any time in Settings › AI features. With it off, the app works as before and nothing is sent for AI. Nothing already sent is kept longer because of this; it is deleted on the schedule below.

Who processes the AI data, and where

We use the following processors, each under a data processing agreement and only to provide the feature you started:

• Google Cloud and Firebase (Google Ireland Limited / Google LLC) — hosting, anonymous sign-in, App Check, push notifications and processing. Your uploads are stored and processed in the European Union (Belgium, region europe-west1). The records that track each job (see below) are stored in Cloud Firestore in the United States (multi-region nam5).
• NVIDIA Maxine — the eye-contact software. It runs on GPU servers we rent from RunPod (below) or, if RunPod is unavailable, inside our own Google Cloud project in the EU region above. NVIDIA does not receive your video.
• RunPod (Runpod, Inc., United States) — provides the GPU servers that run AI eye contact, in data centres in the European Union and the European Economic Area. The server fetches your recording and returns the corrected copy through short-lived signed links (valid for about two hours); RunPod does not keep your video after the job.
• Groq, Inc. (United States) — transcribes the audio for AI captions. If Groq is unavailable, OpenAI transcribes it instead.
• Google Gemini on Vertex AI (EU region above) — translates captions, and writes a script draft if Claude is unavailable.
• Anthropic PBC (United States) — its Claude models write and rewrite script text for Create with AI and AI rewrite.
• RevenueCat, Inc. — holds your AI subscription status and your AI credit balance and its transactions, as described under “Purchases and subscriptions”.

Neither we nor these providers use your recordings, audio, the text you type or the results to train AI models. We use these services through their business APIs, whose terms exclude training on customer data; some of them may keep a request for a short time to detect abuse, as their terms allow.

How long AI data is kept

• Recordings and audio you upload are deleted as soon as the job ends (a clean-up runs every hour), and in any case within 24 hours. An upload that never becomes a job is deleted after 2 hours.
• Results — the corrected video or the caption file — are deleted from our servers 24 hours after they are ready. The app downloads them to your phone first.
• Topics, notes and script text for Create with AI and AI rewrite are passed to the AI provider and not stored or logged by us.
• Job and credit records — which feature you used, how long the recording was, the credits charged or refunded, the job's status and times, and the settings of the job — are kept for accounting, refunds, fair-use limits and fraud prevention. For a captions job that used your script, its settings include that script text. These records are not deleted on a fixed schedule; we delete them when you ask us to (see “Your rights”).
• Usage counters — how many minutes, uploads and script drafts you used per day or month — are kept to enforce the limits in our terms.

Purchases and subscriptions

Payments are handled entirely by Apple or Google. We never see your card details, billing address or full name.

To know whether your subscription is active, we use RevenueCat, Inc. as a processor. RevenueCat receives a randomly generated, app-specific identifier that is not linked to your name or email, together with your purchase and renewal history, the store you bought from, your country and basic device and OS information.

We use this to unlock Pro and OwlPrompter AI, to keep track of your AI credits, to restore purchases on a new device and to see aggregate sales figures. RevenueCat also passes trial and subscription events to Meta so we can tell which of our ads lead to subscriptions — see “Advertising measurement with Meta” below. We do not use it to build a profile of you.

RevenueCat's own privacy policy is at revenuecat.com/privacy.

Product analytics and ad attribution

To learn which features people use and where OwlPrompter can be better, the app sends anonymous usage signals to TelemetryDeck, a privacy-first analytics service that acts as our processor. No personal data is collected: the signals carry no name, email or account, your IP address is never stored, and the identifier that groups your signals is a salted one-way hash that cannot be reversed to you or your device. We use it only for aggregate statistics — never to identify you, build a profile, or track you across other apps or websites.

When you use video mode, background and face effects are powered by the DeepAR SDK. It processes the camera image on your device — no image or video is sent to DeepAR. To verify our licence, count monthly active users and keep its own anonymous usage statistics, the SDK sends DeepAR a randomly generated device identifier together with basic usage events (for example, that an effect was used) and device information such as the model and OS version; part of these statistics is collected through Google Analytics. This identifier is not linked to your name, email or purchases, and is not used for advertising or to track you across apps. See deepar.ai/privacy.

On iPhone, when you install OwlPrompter after tapping an Apple Search Ads ad, we read Apple’s attribution token so we know which campaign led to the install. Apple provides this in a privacy-preserving way; it does not use the advertising identifier (IDFA) and does not require App Tracking Transparency permission.

You can read TelemetryDeck’s own privacy terms at telemetrydeck.com.

Advertising measurement with Meta

We advertise OwlPrompter on Facebook and Instagram. To learn whether those ads lead to installs and subscriptions — and so that Meta can show them to people more likely to find the app useful — the app includes Meta’s software development kit (SDK), set up for ad measurement only. It is provided by Meta Platforms Ireland Limited if you live in the EEA, the UK or Switzerland, and by Meta Platforms, Inc. elsewhere.

When the app is installed and each time it is opened, the SDK sends Meta that event together with an identifier the SDK creates for this installation, your IP address and basic device information such as the model, OS and app version, language and time zone. On Android it also sends your device’s advertising ID, unless you have deleted it in Android’s settings. On iPhone it does not: we never ask for App Tracking Transparency permission, so the advertising identifier (IDFA) is not collected, and Meta measures iPhone campaigns only through Apple’s SKAdNetwork, which reports installs to Meta in aggregate without identifying you.

When you start a free trial, subscribe, renew or cancel, RevenueCat passes that event to Meta with the product, price and currency, the identifier created by Meta’s SDK, your IP address and, on Android, the advertising ID. Your scripts, recordings, chat messages and Delivery Score reports never go to Meta, and the app does not use Facebook sign-in.

Meta uses this information to measure and improve our ads under its Business Tools Terms, and may also use it as described in its own privacy policy at facebook.com/privacy/policy. On Android you can stop your advertising ID from being used at any time by deleting it in Settings → Privacy → Ads.

This website

Our website (https://owlprompter.app) uses Google Analytics 4, provided by Google Ireland Limited and Google LLC, to count visits and see which pages people find useful. It sets first-party cookies (_ga, _ga_*) holding a random identifier, and sends Google the pages you view, the page that referred you, your browser and device type, and an approximate location derived from your IP address — Google Analytics 4 does not log or store IP addresses. The site sends Google no advertising signals: ad storage, ad personalisation and ad user data are switched off. What you type or record in the site's browser tools stays in your browser and is not sent to us or to Google.

In the EEA, the UK, Switzerland and Türkiye, no analytics cookie is set until you accept in the cookie banner; until then Google receives only cookieless signals that carry no identifier. Elsewhere analytics is on by default. You can change your choice at any time with “Cookie settings” at the bottom of every page. Website analytics data is kept for at most 14 months. Google’s own privacy policy is at policies.google.com/privacy.

Support chat

Settings offers a chat with our support team, run for us by Chatist as our processor. Nothing is sent to Chatist until you open the chat for the first time. From then on it receives the messages you write, an identifier it creates for this device, and the device details a support team needs: device model and name, OS version, app version and build, screen size, battery level, language and time zone. We do not send your name, email address or purchases, and the chat is not used for advertising or to track you. If you allow notifications, Chatist also receives this device's push token (on Android issued by Google's Firebase Cloud Messaging) so it can tell you when support replies.

We use your messages only to answer you. To have your chat history deleted, write to support@techinn.solutions. If you would rather not use the chat, you can reach us at the same address.

Live streaming to other services

When you start a broadcast, your audio and video are sent directly from your device to the RTMP server you chose — YouTube, TikTok, Facebook or your own. The stream does not pass through any system of ours, and we receive neither the content nor any record that a broadcast happened.

Once your video reaches that platform, that platform's terms and privacy policy govern it. Nothing in this policy limits what they may do with it.

The server URL and stream key you enter are stored encrypted on your device — in the Keychain on iOS and behind an Android Keystore key on Android. They are sent only to the streaming server you pointed them at.

Connecting a YouTube account

OwlPrompter may offer to sign you in with Google so it can create a broadcast on your YouTube channel for you. This is optional, is never required to go live, and is switched off in the current release — you can always paste a stream key by hand instead.

If you do use it, the sign-in happens in your system browser and we receive an access token from Google that lets the app create and bind a broadcast on your channel. The token, and a refresh token if Google issues one, are stored encrypted on your device. We do not read your videos, your subscribers or anything else on your account, and no token is ever sent to us.

You can revoke this access at any time at myaccount.google.com/permissions.

Permissions the app asks for

• Camera — to show the preview and record video. Only in video and live modes.
• Microphone — to record audio and to follow your voice while you read.
• Speech recognition — for voice-tracked scrolling and automatic subtitles.
• Notifications — on iPhone and iPad, to tell you when an AI job you started is ready (asked when you start your first one); on Android, to display the ongoing notification that lets a broadcast keep running while the app is off screen.
• Display over other apps (Android) — for the floating prompter overlay.
• Files — read only, and only for the document you pick when importing a script.

Every one of these is requested in context, at the moment the feature needs it, and the app remains usable if you decline.

What we do not do

We do not sell personal data, we show no ads inside the app, we do not build profiles of you, and we do not use your recordings, your scripts or anything you send to the AI features to train AI models. On iPhone we never ask for App Tracking Transparency permission and do not collect the advertising identifier (IDFA). The one place data is shared for advertising is the Meta measurement described above: install, app-open and subscription events, with an identifier created by Meta’s SDK and, on Android, the advertising ID, so we can tell which of our Facebook and Instagram ads work. Everything else we measure is anonymous product analytics through TelemetryDeck (no personal data, no IP stored) and Apple’s own Search Ads attribution, which tells us which campaign led to an install without identifying you. Our website uses Google Analytics, as described under “This website”.

Children

OwlPrompter is a tool for presenters and creators and is not directed at children under 13 (or under 16 where local law sets that age). We do not knowingly process their personal data. If you believe a child has provided us with data, write to us and we will delete it.

Why we are allowed to process this data

Under the GDPR we rely on the performance of our contract with you (Article 6(1)(b)) to verify your purchase and unlock the features you paid for, and on our legitimate interests (Article 6(1)(f)) in understanding aggregate, non-identifying sales figures and in preventing fraudulent purchases. Under the KVKK the corresponding grounds are Article 5(2)(c) and 5(2)(f). We answer support chats on the basis of our legitimate interest in helping you when you ask (Article 6(1)(f); under the KVKK Article 5(2)(f)). Website analytics cookies are set on the basis of your consent where the law requires it (Article 6(1)(a); under the KVKK Article 5(1)), and otherwise on our legitimate interest in knowing how the site is used (Article 6(1)(f)). We measure our advertising with Meta on the basis of our legitimate interest in knowing which of our ads work (Article 6(1)(f); under the KVKK Article 5(2)(f)).

When you start a cloud AI feature, we process what it sends to provide the result you asked for, which is the performance of our contract with you (Article 6(1)(b); under the KVKK Article 5(2)(c)), and we ask your permission in the app before the first upload of each feature. Fair-use limits, the credit records and the once-per-device free allowances rest on our legitimate interest in preventing abuse and keeping accurate accounts (Article 6(1)(f); under the KVKK Article 5(2)(f)). Your face appears in an eye-contact recording, but we do not use it to identify you or anyone else.

Your rights

Depending on where you live, you may have the right to access the personal data we hold about you, to correct it, to have it deleted, to restrict or object to its processing, to receive it in a portable format, and to withdraw consent where processing is based on consent.

In practice the data we can act on is what our payment processor holds against your anonymous identifier and, if you have used the cloud AI features, the job and credit records linked to your anonymous user ID; everything else is on your device and outside our reach, and your uploads are deleted within 24 hours anyway. Write to support@techinn.solutions and we will respond within 30 days; because there is no account, we may ask for something that lets us find your records, such as a purchase receipt. If you have used the support chat, we can also access, export or delete that conversation for you. Data Meta receives for ad measurement is held by Meta: you can exercise your rights over it with Meta directly, and on Android you can delete your advertising ID at any time.

You also have the right to complain to your local supervisory authority — in the EU your national data protection authority, in the UK the ICO, and in Türkiye the Kişisel Verileri Koruma Kurumu.

How long data is kept

Purchase records are retained by our payment processor for as long as your subscription is active and afterwards for as long as tax and accounting law requires. Data for the cloud AI features is kept as described under “How long AI data is kept”. Everything held on your device is kept until you delete it or remove the app. Support chat conversations are kept for as long as needed to handle your request, and deleted when you ask us to. Meta keeps ad-measurement data according to its own retention policy.

International transfers

Our payment processor and the app stores operate globally, so the limited purchase data described above may be processed outside your country, including in the United States. Where that involves a transfer out of the EEA, the UK or Türkiye, it is covered by the European Commission's Standard Contractual Clauses or an equivalent lawful transfer mechanism. The same applies to the support chat, whose provider may also process your messages outside your country. Google may process website analytics data in the United States; Google LLC is certified under the EU–US Data Privacy Framework. Meta may process ad-measurement data in the United States and other countries, under the transfer mechanisms described in its privacy policy. For the cloud AI features, your uploads are stored and processed in the EU (AI eye contact on servers in the EU/EEA), but job records are stored in the United States and transcription (Groq, or OpenAI as a fallback) and script writing (Anthropic) take place in the United States. These transfers rely on the European Commission's Standard Contractual Clauses or, where the provider is certified, on the EU–US Data Privacy Framework, and under the KVKK on the safeguards of its Article 9.

Changes to this policy

If we change what the app does with data, we will update this page and move the date at the top. Material changes will also be described in the app's release notes. Continuing to use OwlPrompter after a change means you accept the updated policy.

Contact

TECHNINN SOLUTIONS BİLİŞİM TEKNOLOJİLERİ A.Ş.
İnkılap Mah. Sokullu Cad. No:5 Ümraniye / İstanbul, Türkiye
support@techinn.solutions

OwlPrompter · TECHNINN SOLUTIONS BİLİŞİM TEKNOLOJİLERİ A.Ş. · https://owlprompter.app