AI Characters & Your Data
Effective November 1, 2026 · Version 4.0
How AI powers Wibu's characters, what data AI features use, what stays on your device, and what AI cannot do.
Introduction
Wibu's characters are anime characters powered by AI. This notice explains in plain words how that works and what data it uses. It adds to the Privacy Policy and replaces the Wibu 3.0.0 "AI Disclosure & Data Sharing" notice.
In short
- Every Wibu character is an AI character. The App shows this in the character's room.
- To make characters reply, speak and create content, we send the data needed to third-party AI service providers.
- Those providers may use the data only to provide the service to us.
- In the App, your voice in calls is turned into text on your device.
- You can see and delete what characters remember about you.
1. Characters are AI
1.1 Characters are fictional anime characters, written and voiced by AI. They are not real people, and they are not conscious.
1.2 Characters stay in character. They have their own personality, tastes, daily routine, moods and small story arcs, and they may call themselves an anime character. All of this is fiction. Their "life" is generated automatically and is the same story for every fan, shared at different bond levels.
1.3 Characters may message you first, for example to ask how something went. These messages are limited, respect quiet hours, and can be muted per character or turned off in Settings → Notifications.
1.4 Characters are designed never to pressure you to pay, send gifts or come back, and never to use guilt or jealousy. If you see that happen, please report it.
2. What AI features use
| Feature | What is sent to AI service providers |
|---|---|
| Chat replies | Your message, the character's persona, recent conversation turns, relevant memories, and your nickname or first name if the character uses it |
| Calls | The text of what you said (turned into text on your device in the App), the same context as chat, and the reply text to be spoken |
| Voice messages and voice lines | The text to be spoken and the character's voice |
| Selfies, clips, video messages and avatars | The character's look and prompt; for a character's look, any photo you upload |
| Songs | The song request and the name the character uses for you |
| Voice cloning (creators) | Your voice recording and the cloned voice |
| Safety checks | Messages, images, posts, voice lines and other content, to detect banned material |
| Memory | Your chats, to pick out facts and moments the character should remember |
| Profile photos | Your photo, to check it is safe before others see it |
Some features, such as daily life plans, trending topics and event planning, use AI without your personal data.
We use different AI models for different features and plans, for example a more advanced model for PRO "deep talk" messages. Our main AI providers currently include Google, Anthropic and OpenAI. Providers may change over time.
3. What stays on your device
| Action | Where the data goes |
|---|---|
| Speaking in a call in the App | Turned into text on your device. We receive only the text. |
| Speaking in a call on the web | Turned into text by your browser's speech feature, under your browser maker's terms |
| Recording a clip to share | Made on your device. A copy may be processed on our servers to create a video file, and is deleted after a short time. |
| Browsing the feed, reading posts | No AI provider is involved |
4. Consent and control
4.1 Before you first use AI features, we ask you to agree to this processing. This is a one-time screen before your first chat.
4.2 Memories. Characters show what they will remember ("{name} will remember: …"). In Settings → Character memories you can delete one memory, all memories for one character, or everything.
4.3 Proactive messages. Mute a character from the chat menu, or turn off character notifications in Settings → Notifications.
4.4 Voice clones. You can ask us to delete a voice you cloned. We also ask the provider to delete it.
5. Training
- We do not sell your data to AI providers.
- We do not license your chats, characters or creative content for training public or third-party AI models.
- Our contracts and settings with AI service providers limit them to processing your data to provide the service, and not to train their general models on it.
- We may use de-identified and aggregated data to improve our own prompts, safety systems and features.
- Content flagged by safety systems may be kept and reviewed by people for safety.
6. What AI cannot do
- AI can be wrong. Replies may be inaccurate, inconsistent, out of date or unexpected. Characters may talk about real anime, games or music, and may get facts wrong.
- Not professional advice. Wibu is not a therapy, medical, legal, financial or crisis service. If you are in crisis, contact a professional or local emergency services.
- Not a real relationship. A character cannot replace real people or real support.
- Not unique. Similar output may be made for others.
7. International transfers and retention
AI service providers operate in several countries, which may include the United States, Singapore, Japan, the European Economic Area and Hong Kong. Where required, we use safeguards such as Standard Contractual Clauses. Providers keep request data only as long as needed to process it and for their own abuse monitoring. Our own retention periods are in section 8 of the Privacy Policy.
8. Age
AI features are for users aged 16 and over. Until you turn on 18+ mode in Settings and confirm that you are 18 or over, the content filter stays on and 18+ characters and features are hidden.
9. Changes and contact
We update this notice when we change how AI features use data. Where a change means sending your data somewhere new in a way that needs consent, we ask first. The current version is always at wibu.ai/ai. The old address wibu.ai/ai-disclosure redirects here.
GEN A LIMITED
Workshop 1104, Block C, Hong Kong Industrial Centre
489–491 Castle Peak Road, Lai Chi Kok
Kowloon, Hong Kong
support@gena.studio
Questions: support@gena.studio