
Privacy policy
- BrainDump has no account, no ads, and no analytics or crash-reporting tools. It doesn't track you.
- Your books, reading positions, marks, notes, reading stats and settings are stored on your device.
- The AI features are off until you tap Allow. After that, BrainDump sends the text of the books you add, your questions, your marks and notes, and your library to our server. Our server passes them to AI models (DeepSeek, or others through OpenRouter) and, for some questions, to web search (TypeSafe, Tavily).
- You can turn the AI features off at any time in Settings › Book intelligence › Turn off AI features.
Who we are
BrainDump (“we”, “us”) makes the BrainDump app for iPhone and iPad. If you have a question about this policy or your data, write to support@alpha-aal.com.
What stays on your device
These are stored only on your iPhone or iPad:
- your books (the EPUB files you add, and the free classics that come with the app)
- your library and collections
- reading positions and reading sessions
- marks, notes and the words you look up
- reading stats and settings, including your AI permission
- copies of AI results: chapter summaries and past Ask conversations
The home-screen widgets show a small snapshot of your library (titles, authors, covers and progress). That snapshot stays on your device too, and so do books you share to BrainDump from other apps.
There is no BrainDump cloud sync. Your device's own backups (iCloud Backup or a computer) include BrainDump's data. Apple handles those backups under its own terms.
The AI features
Ask, recaps, character and place cards, Clarify, book memories and reading suggestions use AI models. They are off until you allow them, either in the welcome or in Settings › Book intelligence. If you don't allow them, nothing is sent and the reader works fully without them.
What is sent once you allow them
| What | When, and why |
|---|---|
| Text from your books | When you add or open a book, its full text goes to our server, chapter by chapter. The server builds an index from it (chapter summaries, and notes on characters and places) so answers can draw on the book. To work out what kind of book it is, BrainDump also sends short excerpts (the opening, a passage from the middle, the copyright page and the table of contents), and topic tags from your other books. |
| Your questions | When you use Ask, with the earlier turns of the conversation, the page you're on, any text you selected, matching passages from the book, and how far into the book you are. |
| Your marks and notes | Highlighted passages, with your notes on them, when they help an answer, a book memory or a suggestion. |
| Your library | For reading suggestions, when you near the end of a book, finish it, or ask for suggestions on Today. BrainDump sends up to 400 books from your library: title, author and ISBN; whether you're reading it, finished it, want to read it, set it aside or gave up on it; how far you've read; and when. It also sends your collections (names and the titles in them), up to 30 words you've looked up, your device's languages, and earlier suggestions with whether you kept them. Once a day, it sends the titles, authors and ISBNs on your Up next list to check for new releases. |
| Other details | Character and place cards send the name you tapped and the text on your page. Clarify sends the passage you selected. Recaps send how many days it's been since you last read. Every request carries the app's version and a random request number. |
| An install number | A random number made when you install BrainDump. It goes with suggestion requests so our server can return recent suggestions for this install instead of making new ones. It isn't connected to your name, your Apple Account or your device's identifiers. |
Who receives it
| Recipient | What it receives |
|---|---|
| Our server alex-dump-backend.fly.dev, hosted by Fly.io in Singapore |
Everything listed above. |
| DeepSeek the default AI model |
What each feature needs: book text, questions, marks and notes, and your library for suggestions. DeepSeek's privacy policy says it processes data in the People's Republic of China. |
| OpenRouter which forwards to other AI model providers, such as Google (Gemini) |
Character and place cards, and Ask when another model is chosen or DeepSeek is unavailable. We tell OpenRouter to use only providers it lists as not collecting the data. |
| TypeSafe | For questions about non-fiction or about books set in the real world: your question and the book's title and author, to decide whether a web search would help. |
| Tavily web search |
If a search would help: your question with the book's title and author. For suggestions: searches such as “books like title by author”. |
| Book catalogs Google Books, Apple, Open Library, Wikidata, and Douban for Chinese titles |
Book titles, authors and ISBNs only, to check that suggested books exist and to find covers, series and release dates. Your marks, notes and library don't go to them. |
Our server makes these requests itself, so these companies see our server, not your device or its IP address. The one exception: covers for suggested books load on your device straight from those catalogs' image servers. Each company handles what it receives under its own privacy policy.
What our server keeps, and for how long
Our server doesn't store your questions or build a profile of you. Here is what it keeps:
- Book text: deleted as each chapter is indexed, and within 48 hours at most.
- Titles of books being indexed and their chapter titles: until 7 days after indexing finishes.
- Book indexes: AI-written chapter summaries, short quotes, and notes on a book's characters, places and series. An index belongs to the book, not to you: anyone who adds the same edition uses the same index. Indexes have no set end date; some parts are deleted 90 days after they're made, and the rest is kept.
- Ask answers: up to 30 minutes, so an answer cut off by a lost connection can resume.
- Some other AI results (character and place cards, book memories, book-type checks): up to 24 hours.
- Suggestions: your library, marks and notes are deleted as soon as the suggestions are ready (or after about an hour if something fails). The suggestions themselves, the title of the book they follow and your install number are kept for 30 days.
- Catalog and search results: 30 days.
- Usage counts: how many requests were made, tokens used and how long they took, with no content. These are kept for running the service.
Our server's logs hold technical details, such as which feature was used, how long it took and how many tokens it used. They don't hold your questions, book text, titles, notes or IP address. Our host, Fly.io, may keep network logs under its own policy.
The server's storage is backed up daily and each backup is kept for 5 days, so deleted data can remain in a backup for up to 5 more days.
Turning the AI features off
Go to Settings › Book intelligence and tap Turn off AI features. BrainDump stops sending at once and cancels any indexing in progress (one cancel message goes to our server). You can allow them again in the same place.
Turning the AI features off doesn't delete copies of AI results already on your device, and it doesn't delete what our server already holds. That expires on the schedule above.
Deleting your data
- On your device: deleting a book removes it, its marks and its reading stats. Deleting the app removes your books and everything else BrainDump stored about your reading.
- On our server: with no account, we can't tell which stored data came from you, and most of it expires on the schedule above. To have something removed sooner, such as the index of a particular book, email support@alpha-aal.com with the book's title and we'll delete what we can find.
Other connections the app makes
- Open Library: when you search for a book to log one you finished elsewhere, your search goes from your device to Open Library (openlibrary.org, run by the Internet Archive). Covers for those books load from covers.openlibrary.org. This doesn't depend on the AI permission.
- Google and web links: “Search” on selected text opens Google in an in-app browser, and Google receives that text. Links in books, and the web sources shown under some answers, open in the same in-app browser. Those sites handle your visit under their own policies.
- Manga sources you add: BrainDump comes with no manga sources. If you add a source repository, the app connects to it and to the sites its sources read from, sending your searches and the titles you open, and keeping those sites' cookies on your device. Other people run those sites under their own terms. When you open the manga section, the app asks our server for a list of sources we've turned off; that request carries nothing about you. If you report a source, the report (the source's name and address, the reason and any note you write) goes to our server.
- Apple: the dictionary, translation and dictation use Apple's services and on-device features. If you've chosen to share analytics with app developers in your device's settings, or you test BrainDump through TestFlight, Apple may give us crash reports and usage statistics. Apple provides those; BrainDump has no code of its own that does this.
Advertising and tracking
BrainDump shows no ads. We don't track you across other companies' apps or websites, and we don't sell or share your data for advertising.
Children
BrainDump isn't made for children under 13. If you believe a child has sent us personal information, write to us and we'll delete what we can find.
Your rights
Depending on where you live (for example under the GDPR in the EU and UK, or in California), you may have the right to access, correct or delete your personal data, to object to how it's used, and to withdraw your consent. For the AI features we rely on your consent; for everything else, on providing the features you ask for.
With no accounts, almost all of your data is on your own device, under your control. For anything else, email us. You can also complain to your local data protection authority.
Our server is in Singapore, DeepSeek processes data in China, and the other companies above may process it in the United States or elsewhere.
Changes to this policy
We'll post any change on this page with a new effective date. If a change means BrainDump will send a new kind of data to AI, the app will ask for your permission again before sending it.