Privacy

• Account: your email address, the display name you give us, and, if you sign in with Google or Apple, the identifier they return. Passwords are hashed by our authentication provider and never visible to us.
• Your books: everything you write or accept — premise, characters, setting, outline, chapters, versions, notes — and the proposals the AI made along the way, including the ones you rejected, so the record of what was suggested and what you chose stays complete.
• Usage counters: how many generations and Assistant messages you have used this month, and words written per session, for plan limits and the “words this week” figure.
• Billing: your plan and, on a paid plan, the Stripe customer and subscription ids. Card numbers are entered on Stripe’s pages and never reach our servers.
• Share links: for a chapter you share, the link token, view count and any reactions a reader leaves. Readers are not identified.
• Preferences: your theme, and whether storyboard sketches are on.

We set two kinds: the session cookies our authentication provider needs to keep you signed in, and two small cookies that remember whether you chose the dark or light theme so the page paints correctly before it loads. Both kinds are strictly necessary, so there is no cookie banner; each one is listed on the cookies page. There are no advertising or cross-site tracking cookies. Our host records standard request logs (IP address, user agent, path, timing) for security and debugging and keeps them for a limited period.

The Service runs on other companies’ infrastructure. Each one sees only what its job requires:

• Supabase — database, file storage and authentication. Your account and books live here.
• Vercel — hosting for the website and the application servers.
• Anthropic — the language models behind proposals. When you ask for a proposal, the relevant parts of your book are sent to generate it. Anthropic’s API terms prohibit training on this data.
• Stripe — payments and subscription management, on a paid plan.
• Google / Apple — only if you choose to sign in with them, and only the identity they return.
• A web-search provider — only for the research feature, and only the query you type.

We do not share your data with anyone else except as required by law.

• We do not train models on it, and we do not permit our providers to.
• We do not read it except to support you at your request, or to investigate abuse reported to us.
• We do not publish it. A chapter is visible to others only through a share link you create, for as long as you leave it open.

• Export: every plan can download a whole book as plain text from “Read as a book”; paid plans add ePub, Word and other formats.
• Deletion: deleting a book removes it and its proposals. Deleting the account, on the account page under “Export and delete”, cancels any subscription and removes everything at once. If you cannot sign in, write to privacy@draftbloom.com from the address on the account and we complete it within 30 days.
• Access and correction: your name, password and preferences are editable on the account page; the export holds everything else; for a correction we cannot make ourselves, email us.

If you are in the EU, UK or another jurisdiction with data-protection rights, the rights of access, rectification, erasure, portability and objection apply and you can exercise them at the same address.

We keep your data for as long as your account exists. Backups roll off on a fixed schedule after deletion. Data is encrypted in transit and at rest by our providers, and every database row is scoped to the account that owns it. No system is perfectly secure; if we learn of a breach affecting you we will tell you without undue delay.

The Service is not directed at children under 16 and we do not knowingly collect their data.

When this page changes in substance we update the date at the top and, for account holders, say so by email. Questions about your data and requests under data-protection law go to privacy@draftbloom.com; a security concern goes to security@draftbloom.com.