Suyai
Privacy Policy - Photo AI
Jonathan Benavides Vallejo ("Owner", "we", or "us") provides Photo AI (formerly Banana Edits, the "Service") as a commercial app. This Service is intended for use as-is.
This Privacy Policy explains how personal information is collected, used, and disclosed when you use the Service. Cloud AI processing is subject to the permission described below.
The terms used in this Privacy Policy have the same meanings as in our Terms of Use unless otherwise defined herein.
Information Collection and Use
When you use cloud AI features, the selected photos and instructions you submit are processed to generate or edit images, as described in "AI processing and local photo tools" below. We also process subscription, usage, and diagnostic information to operate the Service, manage access, prevent abuse, and resolve support requests. If you contact us, we receive your email address, message, and the information you choose to include.
App configuration
Each time you open Photo AI, whether or not you use cloud AI, the app connects to Firebase Remote Config (Google). It checks for the app's configuration, such as which features and safety checks are switched on, and stays connected to receive updates. For this, Firebase gives your installation a random Firebase installation ID and receives technical information such as the app version, operating-system version, language and region settings, time zone and IP address. We use it only to deliver the configuration. Firebase's components also send Google limited diagnostic data, which Google uses to maintain and improve Firebase.
Cloud AI requests
Updated app versions send each cloud AI request through our Firebase service (Google Cloud Functions in the europe-west1 region, Belgium). Firebase App Check uses Apple's App Attest to help verify that requests come from genuine app installations. With each request, the app sends its RevenueCat app user identifier, a random identifier that RevenueCat assigns to your installation, so the service can ask RevenueCat whether your subscription is active. From it, and from RevenueCat's identifier for your subscription, the service derives pseudonymous subscriber codes that it uses to prevent abuse; it does not store the identifier itself. No account or sign-in is required. Our service does not store your photos, instructions or generated images: it processes each request and returns the result to your device.
Safety review
Before an image is created, every request is checked against our content rules:
- On your device, the app checks typed instructions against a list of blocked terms and analyzes selected photos: an explicit-content score and the number of faces and people it detects.
- Where available, the app then sends the instruction, these on-device photo signals and, when the model supports images, small, reduced-size copies of the photos to Apple Private Cloud Compute, which runs on Apple's servers, to check the request.
- Our Firebase service then sends the instruction and small, reduced-size copies of the photos to OpenAI (or Google Gemini, when we select it) to check the request against our content rules. This review may rewrite the instruction so that it follows the rules, for example to keep people from different photos side by side, or it may refuse the request.
- Google Gemini then creates or edits the image from your photos and the original or rewritten instruction.
- Before the result is saved, the app checks it on your device.
- When Apple Private Cloud Compute objects only to the wording of an instruction you typed, it may propose another wording. The app keeps it on your device and shows it only if our review then refuses the request; it is sent only if you choose to send it as a new instruction.
Service records
Our service logs technical records for each request: the pseudonymous subscriber codes, the app build, whether access comes from a purchase and from the App Store or a test environment, the type of request, the number of photos, the model and image size used, timings and usage counts, the safety review's provider and decision and the content-rule categories involved, and error reasons. These records do not contain your photos, instructions, rewritten instructions or generated images. Google Cloud's request logs also record technical information such as the originating IP address. These logs are kept in Google Cloud Logging under its retention settings (currently 30 days). Counts of requests refused under our most serious content rules are also kept as monitoring data labelled with the subscriber code, so that we are alerted to repeated abuse; Google Cloud Monitoring keeps this data for 6 weeks, and an alert sends the code to us. OpenAI, Google and Apple process the information they receive under their own policies and terms, which may include limited retention to detect abuse.
Older versions and usage monitoring
Older app versions (3.0.1 and earlier) may continue to send the photos and instructions of AI requests through AIProxy to Google Gemini (versions 2.2.0 and earlier also to fal, fal.ai) until that route is retired; they will then require an update to generate images. These versions don't use our server's safety review, content reports or suspension, and they don't add the file-metadata or visible AI labels described in this policy. AIProxy associates requests with a device-related identifier supplied by the app and processes technical request information, such as originating IP address, response status, error information and request metadata, under its privacy policy. Usage monitoring helps investigate failures and abuse. RevenueCat processes an app user identifier and subscription information to manage purchases and subscription access.
Support diagnostics
When you choose Contact Support, the app prepares an email to support containing your RevenueCat ID, the pseudonymous subscriber code of your installation (the same code our service uses in its records), the app version, your device type (iPhone or iPad), operating-system version, language and region settings, your device's identifier for our apps (which older versions used with AIProxy), the image service last used in the current session and the active AI image settings on your phone. Below your message, the email asks you not to delete these details, which help us identify your installation and device and investigate your issue. If your device has no Mail account, the app opens your default mail app or lets you copy these details. You can review and edit the email before choosing to send it. We use the information you send to troubleshoot your request.
Content reports
You can report any AI-generated image from its menu. A report is an email that you send from your own mail app to our support address; nothing is uploaded to our servers. It contains the reason you choose, your optional comment, how the image was made (the type of request, the filter or idea, the instruction sent and the date), the identifiers listed under "Support diagnostics" and, if you leave "Attach this image" on, a reduced copy of the image without its file metadata ("Attach this image" starts off for "Sexual or nude content"). A report about a child never includes the image. We use reports only to review the content and enforce our content rules, which can include suspending AI features for a subscription. We delete reports once they have been reviewed, unless the law requires us to keep or pass on information. A report that concerns the safety of a child may be passed to the authorities (in Switzerland the Federal Office of Police, fedpol; in the United States the NCMEC CyberTipline; otherwise the police of the country concerned), and we keep its text and identifiers (never the image) for up to 12 months or as long as the authorities ask. You can choose to delete the reported image from the app when you report it.
Purchases
RevenueCat processes an app user identifier and your subscription information to manage purchases and subscription access. Our service asks RevenueCat, with that identifier, whether your subscription is active before each AI request. The app shows RevenueCat's paywall and subscription screens, which tell RevenueCat when they are shown or closed. The app also lets RevenueCat collect Apple's AdServices attribution token, which tells us whether an install came from an Apple Search Ads campaign; it is not used to track you across other companies' apps or websites.
Marking of AI-generated images
Images created or edited with cloud AI carry a machine-readable label in their file metadata (the IPTC "Digital Source Type", which says the image was made with generative AI) and Google's invisible SynthID watermark, which Google Gemini adds to every image it creates. The label contains no personal information. It stays in the file when you save it to Photos or share it. When you save or share one of these images, the app also adds a small "AI-generated · Photo AI" label in a corner of the picture; you can turn this off in Settings under Label AI Images. Photos finished with the on-device tools are not labelled.
Enforcing the content rules
If a subscription breaks our content rules, we can suspend its AI features by adding its pseudonymous subscriber codes to a block list in the app configuration described above. The list holds only these codes, and like the rest of the configuration it is downloaded by every copy of the app; a code doesn't show who it belongs to. Our service then refuses AI requests from that subscription or installation. Codes stay on the list until a review shows the suspension was wrong or an authority asks us to remove them. If requests from a subscription are refused repeatedly because they would sexualize or endanger a child, we may report that pattern (the subscriber codes, the dates and the rules involved) to the authorities named under "Content reports" and keep that information for as long as they ask.
Legal bases
Where the EU or UK General Data Protection Regulation or the Swiss Federal Act on Data Protection applies, we rely on these legal bases:
- Cloud AI requests (sending your photos and instructions to the safety review and to Google Gemini) are needed to provide the features you ask for. We send them only after you allow it in the app, and you can withdraw that permission in Settings under AI Processing at any time; this does not affect requests already made.
- Checking your subscription with RevenueCat and managing purchases are needed to provide the subscription you bought.
- The safety review, App Check, our service records and request logs, the pseudonymous subscriber codes, the block list, the app configuration and the review of content reports rely on our legitimate interest in keeping the Service secure, lawful and free of abuse, and in protecting other people, especially children.
- The Apple Search Ads attribution token relies on our legitimate interest in knowing which of our campaigns bring installs.
- We use support emails to answer your request.
- We keep or pass on information to the authorities where the law requires it or where it is needed to protect a child.
Where we rely on legitimate interests, you can object at any time by writing to us.
International transfers
Our Firebase service runs in the European Union (Belgium). Google, OpenAI, Apple, RevenueCat and AIProxy may process information in the United States and other countries. These transfers rely on the safeguards in the providers' terms, such as the European Commission's standard contractual clauses or, where a provider is certified, the EU-US and Swiss-US Data Privacy Frameworks. You can ask us for details of these safeguards.
Retention
Photos, instructions and results stay on your device until you delete them; our service does not store them. Service records are kept for 30 days and monitoring data for 6 weeks, as described above. We keep support and report emails only as long as we need them to answer and resolve your request, and delete them within 6 months after that, unless the law requires us to keep them longer or they concern the safety of a child (see "Content reports"). RevenueCat keeps purchase information under its privacy policy.
Your rights and choices
You can ask us for access to the personal data we hold about you, and ask us to correct or delete it, to restrict it or to stop using it (you can object to any processing based on our legitimate interests). You can also ask to receive it in a portable format. Write to us from Settings > Contact Support, so that your subscriber code is included and we can find the records linked to it (our service records are kept for 30 days), or to the address under "Contact Us". We answer within 30 days. Photos and results stay on your device; delete them in the app's History. You can withdraw permission for cloud AI requests in Settings under AI Processing at any time. You can also complain to the Swiss Federal Data Protection and Information Commissioner (FDPIC, edoeb.admin.ch) or, if you live in the European Union, the EEA or the United Kingdom, to the data protection authority where you live or work (in the UK, the ICO).
Log Data
Photo AI contains no crash-reporting software and no advertising or analytics SDK. The technical information our service and Firebase receive is described under "App configuration" and "Service records", and what RevenueCat receives under "Purchases". If you choose in your device settings to share analytics with app developers, Apple gives us crash reports and app usage statistics that don't identify you.
Cookies
Cookies are files containing a small amount of data, often used as anonymous unique identifiers. Photo AI does not use cookies itself and has no cookie settings. Some links, such as the Terms and Privacy links on the subscription screen and links in the subscription management screen provided by RevenueCat, open web pages inside the app; those websites may set their own cookies under their own policies. This policy and our Terms of Use, on suyai.ch, use no cookies. Other links open in your browser.
Service Providers
We may engage third-party companies and individuals for the following reasons:
- To facilitate our Service;
- To provide the Service on our behalf;
- To perform Service-related services; or
- To help us analyze how our Service is used.
Service providers receive information needed to deliver their services, including cloud AI processing, subscription management, and diagnostics. Their handling of information is described in their applicable policies and service terms.
Google Gemini API terms and data use
Security
We value your trust in providing us with your personal information, and we strive to protect it using commercially acceptable measures. However, no method of electronic transmission or storage is completely secure, and we cannot guarantee absolute security.
Links to External Sites
Our Service may contain links to external sites. If you click a third-party link, you will be directed to that external site. We strongly advise you to review the privacy policies of those websites, as we have no control over their content, privacy practices, or policies.
Children's Privacy
Photo AI is not intended for users under the age of 18. We do not knowingly collect personal information from anyone under 18. Our content rules allow a photo of a child only if you are their parent or guardian or have a parent's or guardian's permission; like every cloud request, such photos go to the safety review and to Google Gemini as described above. If we become aware that we have collected personal information from someone under 18, we will delete it from our records. If you are a parent or guardian and believe your child has provided us personal information, please contact us.
AI processing and local photo tools
With your permission, updated versions of Photo AI send the photos and instructions needed for each cloud request to the safety review described above (Apple Private Cloud Compute where available, and OpenAI or Google Gemini through our Firebase service) and through Firebase to Google Gemini, which creates or edits the image, for cloud AI generation, editing, transformations and filters. Older versions use AIProxy until that route is retired. These requests are processed outside your device. The providers' applicable policies and terms describe how they handle this information. We share data only with providers bound by terms that protect it at least as this policy does.
Local photo-finishing tools operate on your device. You can turn off permission for future cloud AI requests in Settings. Turning permission off does not recall information already submitted for processing.
Changes to This Privacy Policy
We may update our Privacy Policy from time to time. Users are advised to review this page periodically for any changes. Updates will be posted directly on this page.
This update is effective as of 22 November 2026.
Contact Us
If you have questions or suggestions regarding our Privacy Policy, please contact us at:
Jonathan Benavides Vallejo, Gseckstrasse, 8707 Uetikon am See, Switzerland
Email: jonathan.benavides@suyai.ch