rankion.ai
EU AI Act compliant

AI Transparency

How Rankion uses artificial intelligence — in plain language, honestly, and verifiably.

Translation notice — The German version of this document is the legally binding original. This English translation is provided for informational purposes only. In case of any conflict between this translation and the German version, the German text shall govern. View authoritative German version.

What Rankion does — and which AI is behind it

A tool, not a black-box model.

Rankion is an SEO tool. We help our customers make their websites visible in search engines and AI search engines. To do this we use pre-trained language and image models from external providers — we do not develop our own AI models, we use them as a tool.

Anthropic Claude
Content analysis, optimization, chat assistant
OpenAI GPT & DALL-E
Comparison analyses, image generation
Gemini · Perplexity · Grok
Multi-LLM reality check
Black Forest Labs Flux Pro
Image generation, hosted in Germany
Microsoft Azure OpenAI
Enterprise alternative for GPT models

A complete list with purpose, region and legal basis is available at /en/trust/sub-processors.

When you talk to our chat assistant

EU AI Act Art. 50(1)

Our chat assistant is an AI (Claude Sonnet 4.6 by Anthropic), not a human. We make this clearly visible before every chat session. AI responses can be inaccurate and do not replace legal, medical or financial advice. Your inputs are sent to Anthropic in the United States for processing (zero data retention, EU-US Data Privacy Framework).


When you generate images with Rankion

EU AI Act Art. 50(4)

Generated images contain invisible machine-readable metadata identifying them as AI-generated (IPTC standard DigitalSourceType=trainedAlgorithmicMedia). In our app you also see a visible " AI-generated" label. You may not generate images of real persons with Rankion unless that person has given consent.

What we do not do

We do not deploy Rankion in areas the EU AI Act classifies as "high-risk":

  • No automated decisions about persons (recruitment, creditworthiness, access to education or social benefits)
  • No social-scoring mechanisms
  • No biometric identification or emotion recognition
  • No manipulation or exploitation of vulnerability

The full feature-by-feature self-classification is in the AI Risk Assessment.

Who we are

Rankion is operated by Provimedia GmbH. Responsible party within the meaning of the AI Act: see imprint.
Data Protection Officer: datenschutz@rankion.ai

Status

Created on 3 May 2026.
Version history and full compliance documentation: Trust Center

Cookies: We use necessary cookies for functionality and optional ones for improvements. Details

Necessary
Active
Analytics
Marketing