Team and subject
Create your team. Your 20-minute timer starts when you press the button. Everything you do is saved to the team automatically.
Already have a team? Resume it
Choose your subject
Choose a subject to continue.
Build the prompt
Part 2 of the formula, the verified context, is the facts card above. It is added automatically.
Prompt quality
Your prompt
Use any AI tool you like: copy the prompt, then paste its answer in the next step. Or write the text yourself.
AI's first draft
Humanize the text
Edit the text yourself, or let Claude rewrite it. Eight checks run automatically as you type, including a slop check for empty AI filler.
Tip: mark a verified local detail in [square brackets], for example [local detail: ground floor, room 2].
0 words
Slop watch-list: what to cut
- Filler openers and hype: “In today’s world”, “It is important to note”, “game-changer”, “seamless”, “empower”, “unlock”.
- Vague claims: “a wide range of support” says nothing. Name the service, the day and the hour.
- Repetition: the same opener or word again and again, and three-part lists made only for rhythm.
- Fake specifics: invented numbers, quotes, names or places. Only the brief and your [verified detail] are allowed.
- Test: if you can delete a sentence and the reader loses nothing, delete it.
Humanizing score
Choose the output and build the blocks
Let Claude write the format, or start from your text and edit it by hand.
The app cannot make a picture. Paste a link to a licensed image or one your team made. Check the licence first.
Make it in a tool
Add the result from your tool
Made something in NotebookLM, Canva, Google Flow or Claude Code? Add it here. It appears in the preview and on the wall.
Blocks
Live preview
Final check
This is exactly what the room will see. Six checks score your result.
Final score
EU AI Act self-check
A training check against the AI Act rules that matter for this content. It is not legal advice. Confirm any real publication with your legal contact.
Optional. Write one line: which AI tool, and what a person reviewed. Real organisations keep this log for every product.
What the rules say (checked 18 November 2026)
- Chatbots must tell people they are talking to an AI, unless that is obvious (Art. 50(1)).
- Realistic AI images, audio or video (deep fakes) must be disclosed (Art. 50(4)).
- AI-generated text published to inform the public on matters of public interest must be disclosed, unless a person has reviewed it and holds editorial responsibility (Art. 50(4)).
- Manipulative techniques and exploiting vulnerable people are prohibited (Art. 5).
- General information about migration is not a high-risk system. AI that helps decide asylum, visa or residence cases is (Annex III).
- The Digital Omnibus changes some dates, for example for high-risk rules and for marking older systems. Check the final text.