AI Act transparency kit

From 2 August 2026, anyone using AI on their site has to say so: conversational assistants, generated content, deepfakes, text published on matters of public interest. Declare what you do and this kit prepares the notices, the statement and a summary of the duties that follow.

Legal data updated on 25 July 2026. See every deadline

This tool is technical orientation, not legal advice. Based on what you declare it shows which duties follow: it does not establish whether you are compliant.

Who you are

Organisation or practitioner name

Website address

Contact email

I practise a regulated intellectual profession in Italy (lawyer, accountant, doctor, architect and similar)

Turns on the Italian duty under Law 132/2025 to tell clients which AI systems you use.

What you actually do

A chatbot or assistant talks to people on my site

This covers any system that answers a user directly: support assistant, conversational agent, automated voice on the phone.

I provide a system that generates content (text, images, audio, video)

This is about whoever makes the generative system available, not whoever uses it. If you only use someone else's tool, this case is usually not yours.

I publish deepfakes, meaning content that looks authentic but is generated or manipulated

Images, audio or video resembling real people, places or events closely enough to appear authentic. Clearly fantastical content stays outside.

I publish AI-generated text to inform the public on matters of public interest

Typically articles and news on matters of collective relevance. Commercial or descriptive copy usually falls outside.

What you type stays in your browser and is never stored anywhere. Close the tab and it is gone.

What you need

Tick at least one case on the left to see what you need.

Genuinely free. No sign-up, no email, no limits, no cookies. We don't store the URL you analyse.

No warranty. These tools are provided free of charge and as is. Results are generated automatically, may contain errors, inaccuracies or incomplete assessments, and do not constitute technical, legal or security advice, nor a certification of compliance. The decisions you make on the basis of this report, and their consequences, remain yours: the publisher of this site accepts no liability for damages, losses or missed obligations arising from its use. If in doubt, have the results checked by whoever manages your site or by a professional.

Want to understand what applies to you?

The guide to the transparency obligations explains who has to disclose what, and with which exceptions. The current deadlines separate what starts now from what was postponed.

Frequently asked questions

›Does this tool tell me whether I am compliant?

No, and it could not. It lets you declare what you do and shows which duties follow, with the dates. Establishing whether you are compliant means looking at your actual situation, and that is a job for a professional.

›Where does the data I type go?

Nowhere. Everything happens in your browser: there is no server call, nothing is stored, and closing the tab wipes what you wrote.

›Do I have to disclose the chatbot even if it is obviously a bot?

Article 50 allows an exception where the artificial nature is obvious to a reasonably well-informed person, but the guidance warns against leaning on it. A clear notice before the first interaction costs little and removes the doubt.

›Are the EU icons enough on their own?

No. The icons published by the Commission are free and optional, and they make labelling recognisable and consistent. The duty to label stands, and using the icon does not satisfy it by itself.