Help ↗

Data and privacy

How your reading is handled.

Privacy policy · Updated 24 September 2026

Processed on this device by default. Online features are turned on separately.

FeedEnglish helps you build vocabulary from the English content you read. On-device collection, leveling, and learning records stay on your device by default. Google sign-in is used to verify your account and membership. Cloud sync and online AI are each turned on only if you choose.

1. On-device collection and storage

On approved sites, the extension processes visible English body text you pause to read, and stores words, matching sentences, source links, titles, public author information, reading-topic statistics, and mastery feedback. This is used for your vocabulary, passage review, and learning recommendations. It does not read input fields, passwords, or payment forms, and it is not designed to collect a complete browsing history. Private routes, editing areas, and Incognito contexts are restricted from collection.

YouTube is processed only when you are watching in the foreground and English captions are available; the full caption track is not downloaded. Models, dictionaries, and runtime files ship with the install package. Built-in browser translation language packs are managed by Chrome.

Vocabulary, sources, profile, learning records, cache, and settings are stored in the extension’s local storage and IndexedDB. After you pause collection, no new automatic collection is added. You can export or clear learning data in “Settings → Data and privacy.”

2. Google sign-in and membership

Sign-in uses Google authentication and sends sign-in credentials to the FeedEnglish account service. We store the Google account identifier, verified email, display name, last active time, and trial or membership period, so we can identify the account, provide the service, and handle support requests. Signing in does not by itself upload your vocabulary.

Your first sign-in includes a 7-day trial with no payment method and no automatic charge. Payments are processed on Dodo Payments checkout pages. FeedEnglish receives order and membership status, not full card numbers. Payment details are described in their privacy policy.

3. Optional account sync

Only after you turn on cloud sync does the extension sync vocabulary, saved passages and sources, reading profile, mastery status, learning stats, and related cache to FeedEnglish servers. Device settings and browser activity logs are not part of learning sync.

Pausing sync stops new automatic sync requests; requests already sent may still finish. Avoid editing learning data on several devices at the same time; concurrent edits use the last successful submit. Cloud learning data is kept until you clear it and complete sync, or delete your account.

4. Optional online definitions and AI explanations

“Online definitions and AI explanations” is off by default and must be turned on separately in “Settings → Data and privacy.” After you turn it on, loading definitions in the word list sends the word and topic; opening a word’s details also sends up to the 3 most recent saved sentences and nearby context for sense matching; clicking sentence-level AI explanation sends that word, the selected saved sentence, and nearby context. Requests are forwarded by the FeedEnglish service to Alibaba Cloud Model Studio (Singapore region); if it responds too slowly, fails, or declines the content, they are forwarded as a backup to OpenRouter and its model providers. Either route may involve cross-border transfer. Turning this switch off blocks new online requests; requests already sent may still finish.

Online results are stored on this device. If cloud sync is also on, they sync with the related learning assets. General dictionary definitions may be cached and reused on the server. Explanations of your sentences and context matching use a temporary cache and are not written to a shared persistent cache. Alibaba Cloud, OpenRouter and model-provider data practices also apply. Do not send sensitive original sentences to AI if you do not want a third party to process them.

5. Use limits and service providers

We do not sell personal data, use reading content for advertising profiles, or use personal reading data to train models. Information obtained through Google APIs follows the Google API Services User Data Policy, including Limited Use requirements. Data is used only for the features above, user support, security, and necessary service operations.

Services involved include Google (authentication and the browser), Cloudflare (website, account services and traffic security), Alibaba Cloud (dictionary, language, and AI model services for AI requests you turn on), OpenRouter and its model providers (backup route for AI requests), and Dodo Payments (payments you start). These services may process necessary IP addresses, request times, and error logs. Logs are used for operations and security, and are kept from recording sign-in credentials and request bodies.

6. Deletion, retention, and control

You can turn off collection on selected sites, stop online AI, pause sync, sign out, or export your vocabulary. Signing out or deleting the cloud account does not automatically delete the on-device vocabulary. Account recovery copies may still exist on the device; to remove those copies completely, clear extension storage or uninstall the extension, and export anything you need first.

Deleting your account in “Settings → Account” deletes cloud vocabulary, passages, profile, learning records, and device sync records, and clears email and display name. A minimal account identifier and deletion time are retained so a deleted account cannot be restored from an old device. Necessary audit records for transactions and granted benefits are kept for payment after-sales support; Dodo orders are handled under their rules. If you have an active subscription or a recent unfinished payment request, handle the subscription and payment before deleting the account. Deleting an account does not automatically issue a refund.

Temporary AI processing of your sentences is not written to a shared persistent cache. Beta versions before this policy took effect may have had server-side explanation caches. Historical test caches and backups are limited to server-side administration and are not provided as shared content for public-release users.

7. Policy updates and contact

When data handling changes in a material way, we update this page and the relevant feature notes, and ask for your consent again when needed. For account, data access, or deletion questions, you can use the support link on the FeedEnglish Chrome Web Store listing after publication. The store version has been submitted for review and is not yet publicly installable; installation and account instructions are in Help below.

See help ↗