Otzar is a Hebrew vocabulary app: you look up Hebrew words, save them, and practice them. This page explains exactly what the app sends off your device and why. It is short because there is not much to tell.
When you look up a word, the app sends the word itself and timing metadata to the shared Otzar dictionary server. Looked-up words are stored in that shared dictionary so lookups are fast for everyone.
When you scan a word with the camera, the app also sends the single line of printed text the word appeared in. That surrounding line is used to work out which meaning of an ambiguous word you are looking at, and it is stored in the shared dictionary service so future lookups of the same word can be told apart more reliably. Stored lines are kept with the word they belong to — never with your device identifier — so they are not linked to you.
Requests carry an anonymous per-device identifier — a random string created on your device. It is used only for rate limiting (a per-device daily limit on new words). It is not linked to you: no accounts, no names, no emails, no location.
Words you look up — and, for camera scans, the line of text they appeared in — are sent to Anthropic's API to generate definitions and example sentences. Anthropic's handling of that data is described in their privacy policy.
Your saved vocabulary, practice history, scores, and settings live only on your device. The shared server never sees which words you saved or how your practice is going.
Nothing is sold. Nothing is shared with anyone beyond the two services named above (the Otzar dictionary server and Anthropic's API). There is no advertising and no tracking.
Questions? Write to otzarmilon@gmail.com.