Personal data in a message

The model that answers your agent does not run on your machine. Whatever you send it, leaves.

So while you type, kyte looks at your draft for personal data — names, addresses, phone numbers, account numbers — and says so under the message box: PII detected — click a token to anonymize.

What you do with it

Click a marked piece and it is replaced for the journey. Anna Weber becomes a placeholder; the model sees a person, not a person's name. It answers about the structure, and kyte puts the real values back into the answer on the way in.

Nothing is anonymised behind your back. You choose piece by piece, which is the only way this can work: sometimes the name IS the question.

What is stored

Your message is stored as you wrote it. The anonymisation applies to what travels, not to your record of the conversation.

When it does not appear

The hint only shows while there is text in the draft; sending or deleting it clears the hint with the text.

The bigger question

Where the model runs at all is a separate decision — a workspace can use a model on your own network, and then nothing leaves the building. See which model answers.