Safeguards
What Pluvori doesn’t do
What Pluvori refuses to do is as deliberate as what it does.
You decide the substance
No agent starts on its own. Every step waits for you to act.
Nothing leaves without you
Pluvori publishes nothing and connects to none of your platforms. It never makes up a web address: each channel version carries a link placeholder for you to replace. The only way out is copy and paste.
Nothing is added to your text
Pluvori tells you how it uses AI at the end of the guided tour, without asking you to accept anything. Below the final text, a ready-to-use disclosure note is offered to you, kept separate from your text, with its own “Copy” button. The same note is offered under each channel version, for every channel you haven’t declared private. You’re free to use it, adapt it or ignore it. Each note is written in the language of the text it accompanies.
Pluvori is not a high-risk AI system within the meaning of the EU AI Act. It falls under the transparency obligations set out in Article 50 of that regulation.
Deliberate refusals
No file export: copy and paste is enough, and that’s a choice. An archived folder or article stays archived for good. Your channels and audiences are the ones you name, kept exactly as you wrote them.
Your data
No cookies, no analytics tools, in the app or on this site. Pluvori’s agents run on Claude models, made by Anthropic: your texts are sent to Anthropic for processing. Our privacy policy sets out the terms.