AI Transparency

1. Where We Use AI

Artificial intelligence is part of several FluxSerp features. In each case it produces a draft or an analysis that you review and decide what to do with.

• SEO report assistant — a chat assistant that answers questions about your SEO report.
• Article generation — full articles, titles, meta descriptions and structured data, from your keywords and content plan.
• Image generation — hero and body images for articles, and images for social posts.
• Social content — captions, hashtags and post ideas.
• AI visibility and GEO analysis — we query other AI platforms and report how they describe your brand. This analyses their answers; it does not generate content for you to publish.

2. Which Models We Use

We do not build or train our own AI models. We use general-purpose models from third-party providers:

• OpenAI — text: the SEO assistant, articles, captions and analysis.
• Google Gemini — images for articles, social plan posts and the standalone image generator.

Because these are general-purpose models, their output can be inaccurate, out of date or incomplete. Everything they produce is a draft for you to check.

3. When You Are Talking to an AI

Our in-app SEO assistant is an AI system, and it tells you so before you send your first message. It is not a person, and its answers are generated automatically.

Our support chat widget is staffed by people, not by an AI bot.

4. How We Mark AI-Generated Output

Content produced by FluxSerp carries machine-readable provenance, so that detection tools and platforms can recognise it as AI-generated:

• Images — every image we generate is embedded with XMP metadata carrying the IPTC DigitalSourceType term for media created by a trained algorithm, together with the model used and the time of generation.
• Articles — generated articles record the model and generation time, and publish the same IPTC provenance term in their structured data (JSON-LD).
• Visible disclosure — an article that nobody has marked as reviewed carries a visible "AI-generated" note when it is exported or published through FluxSerp.

You can mark an article as reviewed in the article workspace once a person has checked it and taken editorial responsibility for it. That removes the visible disclosure from later exports and publications. Regenerating the article clears the review, because the text it applied to no longer exists.

5. What This Marking Cannot Do

We want to be precise about the limits:

• We write standard metadata, not a cryptographic signature. Metadata can be stripped by re-encoding, resizing or re-uploading an image on another platform.
• Social captions are copied by you and pasted into other platforms. Once they leave the app there is nothing left for us to mark.
• Text you edit heavily after generation is no longer solely what the model produced, and our records describe the generation, not your edits.

6. Your Responsibilities When You Publish

Under the EU AI Act you act as a deployer of AI when you publish content generated with FluxSerp. That carries obligations we cannot meet on your behalf:

• Disclose AI-generated content where the law requires it on the channels you publish to — especially platforms we do not publish to directly.
• Do not remove the provenance metadata or the disclosure we add. This is also a term of our Terms of Service.
• Decide whether your own review of an article amounts to editorial responsibility before marking it as reviewed.
• Do not use FluxSerp to generate images depicting real, identifiable people. If you produce such an image anyway, labelling it as artificially generated is your responsibility.

7. What We Do Not Do

Some uses of AI are prohibited in the EU. None of them are part of FluxSerp:

• We do not score, rank or profile people. Our scores rate web pages and domains.
• We do not collect or process biometric data, and we build no facial-recognition databases.
• We do not infer emotions from anyone.
• We do not use manipulative or subliminal techniques on end users.

8. Questions

If you have a question about how AI is used in FluxSerp, contact us at contact@fluxserp.com.