Data & Privacy
Most apps treat content you create as their data. You may even sign away rights to your own content when you sign up. We think that’s wrong. Lectr is built so that your library, notes, and quotes stay on your device and are not sent to us.
No Accounts, No Tracking
We don’t know who you are, and we don’t want to. There are no signups, no analytics, and no tracking scripts.
Privacy by Architecture
Lectr uses a local-first model. Your data lives on your device, not on our servers. Where external services are needed, we choose providers with strong privacy commitments and send the minimum data required.
Encrypted Sync
We use Apple’s Private CloudKit Database for syncing. This means your data is stored in your personal iCloud account. We have zero visibility into your records, and if you enable Advanced Data Protection, not even Apple can see them.
iCloud sync is on by default. To disable it, go to Settings > [your name] > iCloud > Apps Using iCloud and toggle Lectr off. Your data stays on your device and nothing is lost.
The “ISBN” Handshake
To help you find and organise books, we are a customer of ISBNdb, a comprehensive book metadata service. When you scan or enter an ISBN, the number is sent to our server which looks it up on your behalf. Only the ISBN is sent; no personal identifiers are attached to the request. We also use the Open Library API as a fallback source for book details.
On-Device Vision
Camera access is strictly for scanning barcodes or photographing pages for quote capture. All image processing happens locally on your device. Nothing is ever uploaded.
Recommendations
The recommendation feature is designed to protect what you actually read. Nothing you type — no tag names, no notes, no quotes — ever leaves your device.
Before any network request is made, your device analyses your reading behaviour locally using Apple’s on-device natural language tools. This produces a statistical summary that includes: recurring terms from your annotations with frequency weights, how your tags cluster together, a sentiment score per topic, how many annotations you typically save per book in each area, and your annotation colour distribution.
Your tag names and annotation themes are then mapped to generic semantic labels before leaving your device. Using Apple’s on-device word embeddings, each term you created is replaced with the closest matches from a fixed vocabulary of broad literary and thematic descriptors. The server receives terms like “sorrow” or “introspective” — never the specific words you chose. This means the AI can reason about the general shape of your reading interests without knowing exactly what you tagged or wrote.
To avoid recommending books you already own, the server generates extra candidates and your device filters out the ones it recognises locally. Your library never leaves your device in any form.
Recommendation requests are stateless. The statistical summary is processed in memory and discarded after each request. Nothing is written to disk or retained.
There are no accounts. Requests are authenticated using Apple’s App Attest and tracked with a pseudonymous identifier that cannot be linked to your identity.
Third-Party Services
Apple iCloud handles library synchronisation if you enable it. Your data is stored in your personal iCloud account under Apple’s privacy terms.
Anthropic provides the AI used for recommendations via its Claude API. The statistical reading summary is included in prompts sent to Anthropic. API prompts are retained by Anthropic for 30 days for safety monitoring and are never used to train AI models.
Cloudflare handles network infrastructure. As with most cloud services, Cloudflare may retain standard connection metadata such as IP address as part of its platform operations.
ISBNdb provides book metadata in response to ISBN lookups. Requests are proxied through our server; ISBNdb never receives your IP address or any reading data.
Open Library is used as a fallback source for book metadata. No reading data is sent.
Zero Lock-In
Your data belongs to you. Export your entire library to standard CSV format at any time.
Questions
If you have questions about privacy in Lectr, you can reach the developer at [email protected].