Last updated: October 24, 2026.
AI Disclosure
1. Introduction
This page explains how UniqAI uses artificial intelligence, which models and providers are involved, what AI-generated content can and cannot do, and what rights you have. It is written to meet transparency expectations for AI systems (including those described in EU AI Act Article 50) in plain language. It does not replace our Terms of Use or Privacy Policy.
2. AI Systems Used
UniqAI does not train its own models. Every chat reply, image and video is produced by a third-party model that you select in the app. The list below reflects the models offered at the date above; the current list is always the one shown in the model picker.
2.1 Chat Assistant
- Models: GPT-4o mini, GPT-5 and GPT-5 Pro (OpenAI); Claude Haiku 4.5, Claude Sonnet 4.6 and Claude Opus 4.7 (Anthropic); Gemini 2.5 Flash Lite and Gemini 2.5 Pro (Google); Grok 4.3 (xAI); DeepSeek V4 Flash (DeepSeek)
- Function: answer questions, write and edit text, help draft prompts for image and video generation
- Notice: the chat is an AI assistant. You are not communicating with a human.
- Limitations: responses may contain inaccuracies, outdated information or biases
2.2 Image Generation
- Models: Flux Schnell and FLUX 1.1 Pro (Black Forest Labs); Nano Banana (Google)
- Function: generate or edit images from your text prompt and, optionally, a reference photo you upload
2.3 Video Generation
- Models: Kling 3.0 (Kuaishou); Wan 2.5 (Alibaba); Veo 3 with audio (Google)
- Function: generate short clips (5–8 seconds) from a text prompt and, optionally, a start image
2.4 How the models are accessed
We access these models through third-party API platforms. Your prompt and any file you attach are sent to the platform and the model maker only to produce the result you requested. See section 4.
2.5 Content Moderation
- Automated checks: prompts are checked automatically before generation, and the model providers apply their own safety filters to inputs and outputs. Blocked requests are not charged.
- Human review: reports of abusive content are reviewed by a person and acted on within 24 hours, as stated in our Terms.
- Transparency: when a request is blocked, the app tells you so and you can rephrase or contact us.
3. Marking of AI-Generated Content
Everything you generate in UniqAI is AI-generated. We do not currently add a visible label, an invisible watermark or C2PA provenance metadata to generated files ourselves. Some model providers may embed their own provenance data in outputs. If you publish generated content, you are responsible for disclosing that it is AI-generated where the law or a platform requires it. We will update this section if we start adding marks to generated files.
4. Data Processing and Third-Party Providers
4.1 What is sent to model providers
When you use an AI feature, the following is transmitted to the API platform and the model maker:
- Text prompts and chat history needed for the current conversation
- Uploaded photos or images used as a reference or start frame
- Generation settings (model, aspect ratio, duration)
We ask for your consent before your content is first sent to an AI provider. We do not send your name, e-mail address or payment details to model providers.
4.2 Provider responsibilities
Model providers and API platforms process this data under their own terms and privacy policies, including their own retention rules. We choose providers that offer API access with business terms, but we do not control their systems. Review the provider's policy if you need to know how it handles API data.
4.3 No automated decisions with legal effect
We do not use AI for automated decision-making that produces legal or similarly significant effects about you. Decisions about accounts (for example a suspension after a report) are made by a person; automated checks only flag content.
5. AI Limitations and Risks
AI-generated content:
- may contain inaccuracies, hallucinations or nonsensical elements
- can reflect biases present in the model's training data
- may resemble existing copyrighted material or real people
- is not guaranteed to match your intent or to be factually correct
- should not be relied on for medical, legal, financial or safety-critical decisions
- may fail or come out low-quality; credits for failed generations are returned automatically
Known limitations include hands, text and complex geometry in images, artifacts or unnatural motion in video, and knowledge-cutoff dates in chat.
6. Prohibited Content
Our moderation and the providers' filters are designed to block, among other things:
- sexual, nude or intimate content
- content depicting violence, self-harm or illegal activity
- deepfakes or impersonation of real people without consent
- any sexual content involving minors
- hate speech, harassment and discrimination
Attempts to generate such content violate our Terms and can lead to account suspension. Report content that slipped through on our Report page.
7. Your Rights
- Explanation: you can ask which model produced a result and why a request was blocked.
- Human review: if an automated check affects your account, you can request review by a person.
- Access and deletion: you can delete your account and the content stored with it from the app's settings, and you have the data rights described in our Privacy Policy.
- Complaint: EU/UK residents may complain to their local data protection authority about how personal data is processed.
To exercise any of these, e-mail support@uniqai.app.
8. Warranty Disclaimer
We make no warranties about AI outputs. Quality and accuracy vary, models can refuse requests because of safety filters, and outputs change as providers update their models. You use AI-generated content at your own risk and should verify it before relying on it.
9. Changes
We may add or remove models and features. Material changes to how AI is used will be reflected on this page with a new "Last updated" date.
10. Contact
Questions about our AI systems: support@uniqai.app. Operator details: see our Terms of Use.