# Manga Enhanced Reader Privacy Policy

Manga Enhanced Reader improves reading on web comic pages. It has no developer account system, ads, telemetry or developer collection server. The developer does not sell user data. The extension does process page content and reading records locally, sync specified data through your browser account, and send necessary information to your configured service when you use AI.

## Reading and local storage

On authorized pages, the extension reads page addresses, comic and chapter titles, image URLs and order, chapter navigation and catalog links, and the page structure and site data needed to parse a chapter. Images load from the site or its image servers; preloading requests nearby images in the current chapter. Those servers process normal network requests, IP addresses and browser-supplied information under their own policies. The extension does not use the cookies API to read cookies, but normal site pages and same-origin reading requests may use the existing login session.

Local storage holds reader preferences, language and theme, chapter positions, reading history, site rules and revision data, AI configuration, cached model lists and recognition records. Reading history includes site and comic/chapter identifiers, titles, chapter URLs, page numbers, page counts and reading times. It covers enhanced-reader use, not your entire browser history. When a comic can be reliably identified, history shows its most recently read position; local per-chapter progress remains separate.

The latest 50 recognition records include time, site/chapter, service origin, model, rule ID, validation counts and failure reasons. They do not store API keys, full page HTML, raw model responses or comic images. The extension does not create an offline comic image library; the browser may cache normal image requests.

## Automatic sync is on by default

Reading history and AI settings use chrome.storage.sync, with automatic sync enabled by default. Synced data includes comic/chapter titles and addresses, reading positions and times, and the AI service URL, model, API key and automatic AI adaptation setting. There is no separate API-key sync switch. The extension adds no application-level encryption to keys in local or sync storage; this is not a password vault. Browser providers control their sync transport, accounts and retention.

Up to 100 recent reading records sync, potentially fewer due to size and browser quotas. Complete history remains local. Images, site rules, model-list caches, recognition logs, site permissions, language/theme and reader display settings are excluded. Devices need the same browser, sync account and extension ID.

You can turn automatic sync off in settings or history. This stops future synchronization; it does not delete existing cloud copies. With sync enabled, successful synchronization propagates reading-history deletions using deletion markers. Clearing an API key or changing and saving the configuration syncs the replacement value when sync succeeds. Immediate removal of browser-service backups is not guaranteed.

## Optional AI services

Automatic AI adaptation is off by default. You may configure a service and run recognition manually, including with a different model. Enabling automatic adaptation allows requests after normal recognition fails. Loading the model list also contacts your service. No model credits are included; the provider's terms determine fees, logging, retention and training use.

AI requests contain the model, rule-generation instructions, schema examples and a size-limited summary of the current page: site origin, chapter URL without query or fragment, title, candidate image selectors/attributes/dimensions/alt text and image URLs without queries, same-origin navigation links and labels, page-count hints, script origins and known protocol/variable names. Titles, paths and visible text can still identify information; the summary is not guaranteed anonymous.

Full HTML, input-field values, cookies and comic image bytes are not sent. Your API key authenticates requests to your chosen provider. It is not given to web-page content scripts or included in rule exports. AI requests omit browser cookies and Referer and do not follow redirects. The service receives the request, authentication data and connection information such as your IP address. HTTP and HTTPS endpoints are accepted; HTTP lacks transport encryption. Use HTTPS for credentials or private content.

The model returns declarative JSON rules using existing bundled engines. Rules are saved only after validation and actual reading checks. Generated JavaScript is not downloaded or executed. Cancellation or page closure cannot retract information already received by the service.

## Retention, deletion and control

Delete individual or selected reading records, or clear reading history. Delete or export custom rules. Update AI configuration or disable sync in settings. Recognition logs retain the latest 50 entries and currently have no individual deletion control. Other local configuration remains until changed, extension data is cleared or the extension is removed. Adjust site access in the browser's extension manager. Uninstalling normally removes local extension data but does not guarantee deletion of browser-cloud or AI-provider copies; manage those through the relevant service.

The developer does not use these data for advertising, sale, credit assessment or profiling unrelated to comic reading, and does not arrange human access to page content. Data is used only for the reading, resuming, synchronization and optional rule-recognition features described here. The extension's use of information complies with the Chrome Web Store User Data Policy, including applicable Limited Use requirements.

## Policy updates

This policy is updated when features or data practices change. The public policy's hosting provider may process ordinary access logs separately from the extension's own network activity.