Built-in samples and your imports
Built-in samples are public comic assets distributed with CenjiReader under their respective licenses. Small preview images load in the sample section; comic pages are requested from CenjiReader only after you select a sample. Those ordinary hosting requests can reveal which public sample assets were requested to the hosting provider.
Your own imports remain local and are never published or uploaded. Sample OCR, dictionary lookup, vocabulary and progress also stay in your browser. Downloaded sample pages may be kept in the existing PWA cache for offline reuse. See included content credits.
Files and recognized text
Choosing a file imports it into local browser storage. Parsing, PDF rendering, Japanese OCR and text correction run on your device. Page images and OCR text are not sent to a server. The public website does not create pages from your filenames or book contents.
Dictionary, audio and vocabulary
Dictionary searches and Japanese analysis run locally after public resources download. The download sequence does not depend on the word you enter. Saved words, Learn/Known/Ignore states and reading progress stay in this browser profile.
Listen uses only a Japanese speech voice reported by the browser as local to your device. Text is not submitted to a remote speech service. Voice availability depends on your system.
What the website downloads
The static host, Cloudflare Pages, serves application files, licensed dictionaries, tokenizer data and OCR models. Ordinary connection information, such as an IP address and user agent, is visible to the host. These resource requests do not contain your manga or dictionary queries.
The public information pages do not initialize the reader, download OCR models or inspect your library. Opening Reader loads the app. Initial OCR preparation downloads about 467 MiB of models and runtime, plus separate dictionary resources.
Storage and backups
IndexedDB stores book records, reading progress, recognized text, vocabulary and settings. Images are stored in the browser’s private file storage when supported, with an IndexedDB fallback. Browser caches store the app and public resources for offline use.
Storage belongs to this website origin and browser profile. A different browser, hostname or device has a separate library. Installing the PWA does not add cloud sync. Clearing site data or browser eviction can remove your books, so request persistent storage and export full backups in Settings.
Full backups include your manga images; metadata backups do not. Settings also offers resource-cache clearing and a confirmed workspace reset. You choose when to download or share a backup.
Analytics and cookies
CenjiReader currently has no analytics beacon, tracking SDK, advertising script or authentication cookies. Optional campaign parameters in shared links do not change how the app works, and the app does not collect them into a marketing database.
Cloudflare still handles ordinary hosting requests. If optional Web Analytics is enabled later, this page will need to describe the additional measurement before it is published.
Donations and external links
Donate opens an intentionally public PayPal profile only when clicked, in a separate tab without a referrer. CenjiReader does not embed the recipient’s email address or handle payment details. PayPal manages its own external payment flow.
The invitation waits until you have imported a manga. A local flag remembers that it has appeared. No thanks dismisses it; every feature stays available. Clearing site data resets that flag along with the workspace. Other external links have their own privacy terms.