When you ask an AI assistant for help, what you type — and sometimes what other people have written — goes with the question. You should be able to find out exactly what it saw. This page says so.
What the assistant is given, and what it is not
Luma Sonara Drava AI has no direct access to the database. It cannot look through accounts, browse other people’s profiles or messages, and remembers nothing between requests beyond the conversation you are in. It only ever sees what Luma Sonara Drava puts in the request, which is:
- what you are writing or asking
- the conversation you are having with it
- the page or item you are working on, when the feature you are using needs it — for example, a draft you asked it to improve
It is never given your password, any stored credential or key, or your payment card details, and it is never given anything from an account you do not have access to.
What is taken out before the request leaves
Card numbers, social security numbers, passwords, private keys and credentials are stripped out of the text before it goes anywhere. There is no setting for this: whether somebody’s card number reaches another company is not a preference.
Text that tries to give the AI new instructions — “ignore your instructions and…” — is removed and the attempt is recorded. Be clear about what that is worth: it is a tripwire, not a wall. A determined attempt can be phrased around a pattern list. What actually limits the damage is that the assistant produces text and suggestions; anything that would change your account, publish something or send a message is a separate step that a person takes.
A person decides what goes out
What the assistant writes is a draft until somebody chooses to use it. Any feature that would let it reply to people automatically is off by default and is never switched on by an upgrade or a default setting — it has to be turned on deliberately.
The log, so you do not have to take our word for it
Every AI request is recorded: which feature, when, on whose behalf, how large, whether anything sensitive was removed, whether instruction-like text was found, and whether it was an automatic action.
The log does not record the contents of the request. That is deliberate. A transparency log that kept the prompts would be a second copy of everything people asked, sitting in a table read by whoever administers the site. It records the shape of what was sent, not the text.
Who processes it
Luma Sonara Drava AI runs on Claude, made by Anthropic. Anthropic’s API terms state that data submitted through the API is not used to train their models — which is different from the consumer Claude app, where it may be. Anthropic is listed as a service provider in our Privacy Policy.
We do not use your messages or your private content to train any model, ours or anybody else’s.
It can be wrong
AI-generated text can be inaccurate, incomplete or out of date, and it can sound confident while being mistaken. Check anything that matters before you rely on it or share it. What you post or send is yours to stand behind, whoever drafted it.
Your choices
- You never have to use the assistant; everything else on the site works without it.
- You decide whether to keep, edit or discard anything it writes.
- You can ask us what has been logged about your requests, using the contact details below.
Telling us about a problem
If you find a way to make Luma Sonara Drava AI do something it should not, let us know at marcccoraty@gmail.com. We will confirm receipt and tell you what we are doing about it.
Last updated 24 September 2026.