Workspace & BillingAI tools

What happens to what you send an AI tool

Where your input goes, who processes it, how long we keep it and what we record about the call.

Updated August 9, 2026

The short version

What you type into an AI tool is sent to us, and from us to a model provider. It is not processed in your browser, unlike almost every other tool on this site.

Who processes it

Model inference currently runs on Cloudflare Workers AI, on Cloudflare's network. Our full list is in Subprocessors, which is the page to check rather than this one, because it is the one we keep current.

What we store

We do not store the text you send or the text the model returns. It passes through the request and is gone when the response is sent.

What we do record, for every AI call:

  • which tool ran, and whether it was the web page or the API
  • the model used
  • how many tokens were consumed — not what they contained
  • what the call cost us and what you were charged
  • the workspace it belongs to, and the time

That record is what your usage page and your invoice are built from. Without it we could not show you why you were charged.

What the model provider does

We use providers on terms that do not permit training on API traffic. That is a contractual position, not something we can technically enforce from our side, and you should treat any third-party model the way you would treat any external service.

What this means in practice

Use AI tools for work you would be comfortable sending to a cloud service. Do not paste credentials, keys, personal data belonging to other people, or anything under a confidentiality obligation that does not permit subprocessors.

For anything you are not comfortable sending, most of our catalogue runs entirely in your browser — those tools are marked as such, and nothing you put into them ever reaches us.

Related