Privacy Policy
This policy explains how AI Talking Photo (“the app”) treats photos, text, and device data. We wrote it for the Google Play User Data policy. The app’s API is hosted on Cloudflare Workers. Provider keys never ship inside the APK.
1. Who we are
AI Talking Photo is a mobile app that turns a photo you choose plus a short line of text into a short AI talking-photo video. Optional face-swap uses a photo of your face on a bundled character image.
The official API is https://talking-photo-api.myuni1001.workers.dev. This page is the public privacy policy for the Play Store listing.
2. Data we process
We do not require an account, email, or phone number to use the app.
- Photos you upload. Face photos you pick from your device to generate a talking clip or a face-swap. These can include biometric-looking face imagery. We process them only to generate the output you requested.
- Text you type. The short line (up to two sentences) that the photo should say, plus the voice and language you pick.
- On-device files you save. Finished videos or images stay on your phone if you save or share them. We do not run a cloud album of your library.
- Local preferences. Language choice is stored on the device (app preferences). We do not sync a user profile.
We do not collect contacts, precise location, SMS, call logs, installed-app lists, advertising IDs, or payment card numbers. V1 has no ads and no analytics SDK.
3. Why we process it
- Create speech from your text (text-to-speech).
- Lip-sync that speech to your photo.
- Optionally swap your face onto a sample character.
- Return the result to the app on your device.
- Keep the API working and stop abuse (a shared app token; request logs at the host).
Legal basis where that language applies: you ask us to generate the clip (contract / requested service) and you confirm you have the right to use the photo.
4. How it is sent (Cloudflare, not the APK)
The Android app talks only to our Cloudflare Worker. It does not call Google, Replicate, or RunPod directly, and it does not embed those API keys.
When you tap Generate, the photo and text go to the Worker over HTTPS. The Worker then calls subprocessors to finish the job:
- Cloudflare — hosts the API, terminates HTTPS, and may retain standard request logs (IP, time, path) under Cloudflare’s terms.
- Google Gemini — text-to-speech for the line you typed.
- RunPod — lip-sync (Wav2Lip) using your photo and the generated audio.
- Replicate — face-swap only, if you use that feature.
Those providers process the files long enough to run the model. We do not sell your photos. We do not use them to train our own public model. We do not build an advertising profile.
We do not operate a permanent server-side gallery of user photos. Generation is request/response. Providers may keep operational logs under their own policies:
5. Photos of children and other people’s faces
Do not upload photos of children (anyone 17 or under). The app is not directed at children under 13 and we do not knowingly process children’s photos.
Only upload a face you have the right to use. Do not impersonate someone or use the output to deceive or harm others. Output is an AI demo, not a real person speaking.
6. Sharing
We share generation inputs only with the subprocessors above, and only to produce your clip. We do not sell personal data. We may disclose information if required by law.
If you use Android share or save-to-gallery, that copy lives on your device or in the app you chose. Those apps have their own policies.
7. Retention and deletion
- No user account to delete.
- Clear the app’s storage / uninstall to remove local clips and preferences.
- Delete saved videos from your gallery if you exported them.
- In-flight jobs exist only while processing; we do not offer a cloud inbox of past uploads.
8. Security
Traffic uses HTTPS. Vendor API keys are stored as Cloudflare Worker secrets, not in the client. The app sends an application token so random browsers cannot use the generate API. No method is perfect; do not upload photos you cannot accept being processed by the providers listed above.
9. International processing
Cloudflare, Google, RunPod, and Replicate may process data in the United States or other countries. If you use the app, you understand your photo and text will leave your device and may be processed outside your country.
10. Your choices
- Do not generate — nothing is uploaded.
- Use sample / public-domain clips only — those images ship in the app.
- Uninstall or clear app data anytime.
11. Contact
Privacy questions: use Feedback in the app Settings, or the developer email on the Google Play listing for AI Talking Photo.
This page: https://talking-photo-api.myuni1001.workers.dev/privacy