Why This Notice Exists
Wibu uses third-party artificial intelligence providers to make characters talk, to generate images, music and video, and to produce voice notes. Doing that means sending some of your data to those providers.
This notice tells you exactly which providers are involved, what each one receives, and what you can do about it. It supplements the Privacy Policy and is published to satisfy Apple App Store Guideline 5.1.2(i), Google Play's Data Safety requirements, and transparency obligations under the GDPR, UK GDPR, the EU AI Act and comparable laws.
- We ask for your consent before the first request to each provider.
- Consent is per feature and per provider — never one global switch, never pre-checked.
- You can withdraw any consent at any time in Settings → Privacy → AI permissions.
- Declining blocks that feature only. The rest of Wibu keeps working.
1The Providers
These are the AI providers we currently use and what each one receives. Providers may change. We will update this page and request fresh consent before sending your data to a provider not listed here.
| Feature | Provider | What is sent | Consent |
|---|---|---|---|
| Character replies in chat | Google (Gemini) | Your message, the character's persona and lore, recent conversation turns, and the character memories selected for that reply | Required, once per account |
|
AI image generation Chat images, feed posts, character art, game art |
Kie | Your prompt, the character's visual description, and any reference photo you upload | Required, first time you generate an image |
| Voice notes | MiniMax | The text to be spoken and the selected voice profile | Required, first time you use voice |
| Custom voice from your own sample | MiniMax | The voice sample you upload, and the derived voice model | Separate consent, asked for on its own |
| Music generation | Our music generation provider | Your prompt and, where you supply them, your own lyrics | Required, first time you generate music |
| Video generation | Our video generation provider | Your prompt and the character's visual description | Required, first time you generate video |
| Safety classification | Our moderation providers and our own systems | Content submitted or generated, for the purpose of detecting prohibited material | Not optional — this is how we keep the platform safe and lawful |
Confirm and insert the exact legal entity names and model SKUs for the chat, music and video providers. Publishing a placeholder where a vendor name is required does not satisfy Guideline 5.1.2(i) or the Play Data Safety declaration, and the in-app consent modals must name the same entities as this page.
2What Stays On Your Device
Not every action sends data to an AI provider. Some features only need permission to read something already on your device.
| Action | Where the data goes |
|---|---|
| Sharing a photo directly in a chat | Stored by us to display in the conversation. Not sent to an image AI provider unless you use it as a generation reference. |
| Setting a profile picture | Stored by us. Not sent to an AI provider. |
| Browsing the feed, reading posts, playing a pre-built game | No AI provider involved. |
All uploads are still checked by our safety systems, and all media you share is subject to the Community Guidelines.
3How Consent Works
3.1 Before the first request
The first time you trigger a feature that sends data to a named provider, we show a consent screen that states the provider, the data type and the purpose. The checkbox starts unchecked and the confirm button is disabled until you check it. There are no pre-ticked boxes anywhere in this flow.
3.2 Managing and withdrawing consent
Settings → Privacy → AI permissions lists every provider relationship with its own toggle and the date you last consented.
- Withdrawing a consent blocks the corresponding request immediately — the feature stops working rather than silently continuing.
- Withdrawing one consent does not affect the others, and does not affect the rest of the app.
- If you re-enable a feature later, we ask for consent again.
- Withdrawal is not retroactive: it does not undo processing that already happened lawfully.
3.3 Custom voices
If you upload a voice sample so a character can speak in a custom voice, we ask for that consent separately from ordinary voice notes, and we use the sample only to produce the voice you asked for, only through the provider named above. Depending on where you live, a voice sample may be treated as sensitive or biometric data, and we handle it accordingly.
Delete it at any time in Settings → Privacy → AI permissions → Voice, or by emailing support@wibu.ai. On request we also instruct the provider to delete it. Absent an earlier request, samples and derived models are deleted within 90 days of the feature last being used, or of account deletion, whichever comes first.
Only upload your own voice. Uploading someone else's voice without their consent breaches the Community Guidelines.
4Training & Model Improvement
- We do not sell your data to AI providers.
- We do not licence your conversations, characters or creative content for training public or third-party foundation models.
- We contract with providers to process your data only to fulfil the request and not to train their own general models on it. Where a provider's standard terms would allow training, we use the enterprise or no-training configuration.
- We may use de-identified and aggregated interaction data to improve our own prompts, safety systems and product. Where consent is required by law, we ask for it or provide an opt-out at Settings → Privacy → Data Use.
- Content flagged by moderation may be retained and reviewed by a human for safety purposes regardless of that setting.
5What AI Cannot Do Here
- Characters are not people. They are software. They are not conscious, they do not have feelings, and a conversation with one is not a relationship with a real individual.
- Wibu is not a therapy, medical, legal, financial or crisis service. If you are in crisis, contact a qualified professional or your local emergency service.
- AI output can be wrong. It may be inaccurate, inconsistent, biased or unexpected. Do not rely on it for decisions that matter.
- AI output may not be unique. Similar output may be generated for other users, and purely AI-generated material may not be protectable by copyright where you live.
All AI-generated content on Wibu is labelled as such, and creators are required to keep that labelling intact.
6International Transfers & Retention
AI providers operate globally. Processing your request may involve transferring data to the United States, the European Economic Area, Singapore, Japan, Hong Kong or elsewhere. Where required, we apply appropriate safeguards including the European Commission's Standard Contractual Clauses and the UK International Data Transfer Addendum.
Providers hold request data only as long as needed to process it and to meet their own abuse-monitoring obligations, typically no more than 30 days, unless a request is flagged for safety review. Our own retention periods are in section 8 of the Privacy Policy.
7Age & AI Features
AI features are available to users aged 16 and over, in line with the platform minimum. Individual features may carry their own age or regional requirements, which are shown in the App.
We do not knowingly process the data of anyone under 16. See section 1 of the Privacy Policy.
8Questions & Changes
We update this notice whenever we add, remove or change an AI provider. Where a change means sending your data somewhere new, we request fresh consent before doing so. The current version is always at wibu.ai/ai-disclosure.
489–491 Castle Peak Road, Lai Chi Kok
Kowloon, Hong Kong
support@wibu.ai