Privacy & your work

You own every word you write here — loglines, characters, beats, scenes, all of it. We don't claim any rights to it. We don't use it as training data for any AI model. We don't share it with anyone outside the services needed to run the app (listed below).

Your screenplays, characters, beats, scenes, ideas, daily-task progress, session messages, and voice-recording metadata live in a Postgres database hosted by Supabase. Each row is tagged with your user ID, and database row-level security ensures no other signed-in user can read your data — full stop.

• Anthropic (Claude API) — your session messages and current screenplay state are sent to Anthropic to generate the guide's replies. Anthropic's API terms don't allow training on customer inputs by default.
• OpenAI (Whisper) — voice recordings are sent for transcription. OpenAI's API terms also don't allow training on inputs by default.
• Supabase — database and authentication.
• Vercel — hosts the web app.

That's the full list. We don't sell your data, we don't syndicate it, and we don't run analytics that touch screenplay content.

A session cookie keeps you signed in (Supabase Auth). A second cookie remembers which screenplay you were last working on so the app opens in the right place. That's it — no tracking, no advertising cookies.

From the Screenplays page, the “Export your data” button downloads a single JSON file with everything tied to your account — screenplays, characters, beats, scenes, every chat message, the lot. No approval, no waiting.

From the Screenplays page, “Delete my account” permanently removes your user and cascades through every table — your screenplays and everything attached to them are wiped. We don't keep a shadow copy. (Database backups taken by Supabase as part of normal operations may briefly retain rows; those roll over on Supabase's standard schedule.)

Email the operator listed on your invitation. This is a small app, not a faceless platform — if something feels off, flag it.