Privacy policy

Your answers (goal, pattern, belief, worst case, visions, senses, why, who benefits) and the three generated scenes are stored so you can come back to them. They are tied to a random identifier generated in your browser — not to your name, email or account — and only that identifier can read them back.

Single-take voice recordings are held in your browser only for the length of your session. They are never uploaded or stored on our servers, which is why leaving the page loses them.

When a scene set is generated we record category-level tags only — broad themes such as "hiring", "pricing", "delegation" or "board relations" — together with a timestamp, so we can see which kinds of goals the tool is used for and improve it.

These records contain no text you wrote: no goal wording, no pattern, no belief, no worst case, and none of the generated scene text. They also carry no identifier of any kind, so they cannot be linked back to you or to a particular scene set. They exist only as aggregate counts per category and are not reviewable as individual submissions.

Generating scenes sends your answers to our server, which calls a large language model to write the three scenes and returns them to you. Your answers are not used to train models and are not retained by us beyond your own saved scene sets.

Deleting a scene set in the app removes it and its version history. Clearing your browser storage removes the identifier that links you to your sets.

Three Scenes is a goal-and-habit rehearsal practice. It is not therapy, medical treatment, or a crisis resource, and nothing on this site should be treated as clinical advice.