This page explains where CaféRadar uses artificial intelligence, what that means for you, and how we label AI-generated content.
1. Where we use AI
We use AI to power a number of features in the app and on our website:
- Barista AI chat: a conversational assistant for coffee questions and recommendations.
- Drink, bag, and photo analysis: when you scan a drink barcode or coffee bag or upload a photo, AI reads the image and returns an estimated description or tasting notes.
- Menu reading: AI helps turn a photo of a café menu into text.
- Recommendations: café and drink suggestions, and the short explanations shown with them.
- Some notifications: a few notifications are written or personalized with AI.
- Coffee blog: some articles on our blog are drafted with AI assistance.
- Merchant insights: AI helps summarize aggregate, count-only analytics for cafés.
These features are powered by third-party AI providers, currently Google Gemini and Anthropic Claude. Our Privacy Policy explains what data is processed and on what legal basis.
2. AI output can be inaccurate
AI output can be wrong. It may be inaccurate, incomplete, or out of date, and it is for general information only. It is not medical, dietary, nutritional, allergen, financial, legal, or other professional advice. Please do not rely on it for any health or safety decision. If you have an allergy, an intolerance, or a health condition, confirm the details directly with the café or a qualified professional before you act.
3. How we label AI content
When you are interacting with AI, or looking at content that AI produced, we tell you with a clear, human-readable label:
- Barista AI chat shows a notice before your first conversation and a persistent “Barista AI” label while you chat.
- In-app analysis of drinks, bags, and photos is marked with an “AI” or “Estimated by AI” indicator.
- Blog articles that were created with AI carry a notice at the top of the article.
- Notifications that are AI-written are covered by this policy and our Terms rather than a separate label on each message.
4. Human review of blog articles
We do not auto-publish AI articles. Before an AI-assisted article goes live on our blog, a member of the CaféRadar team reads it, checks the facts, edits it, and approves it. A named person on our team holds editorial responsibility for what we publish. Articles reviewed this way carry a notice that they were created with AI assistance and reviewed by our editorial team.
5. We do not create synthetic images, audio, or video
CaféRadar does not generate synthetic images, audio, or video, and we do not create deepfakes. Our AI features read information and produce text. Any photos you see in the app are real photos submitted by users or cafés, or provided through our mapping and places partners.
6. Machine-readable marking
For content published on channels we control and keep, such as our blog, we work to include machine-readable information that marks it as AI-generated, in addition to the notice you can see. For short, temporary in-app text and notifications, reliable machine-readable marking is not yet a mature standard, so for now we rely on the visible labels described above. We will adopt a machine-readable method for that content as clear standards become available.
7. Contact us
If you have questions about how we use AI, contact us at:
© 2026 CaféRadar™. All rights reserved.