Gemini Reading Assistant (the "Extension") is a Chrome extension that adds a reading layer — navigation, search, citation, export, and screenshot tools — to conversations on gemini.google.com. This policy explains exactly what data the Extension touches, where it is stored, and what (if anything) leaves your device.
Short version: Everything stays on your computer except a single license-verification call to Gumroad when you activate a Pro license.
| Data | Where it is stored | Purpose |
|---|---|---|
| Extension settings (which modules are enabled, Gemini plan selector, custom prompts) | chrome.storage.local on your device |
Remember your preferences between sessions. |
| Citation cards and memory pins (text you explicitly save) | chrome.storage.local on your device |
Let you reuse saved passages inside Gemini. |
| Conversation snapshots & journals (auto-save of visible messages) | chrome.storage.local on your device |
Restore context and enable Markdown / TXT / JSON export. |
| Pro license data (license code, verification timestamp, purchase email returned by Gumroad) | chrome.storage.local on your device |
Unlock Pro features locally without repeated verification. |
| Screenshot images (Ctrl+Shift+X feature) | Never stored. Downloaded directly to your device via the browser's native download mechanism. | Let you save parts of a conversation as PNG files. |
The Extension makes exactly one type of outbound network call:
https://api.gumroad.com/v2/licenses/verify.
Gumroad returns whether the license is valid and (optionally) the purchase
email so the UI can show the account owner. No conversation content is
ever transmitted.
The Extension does not use any analytics, telemetry, crash reporting, advertising SDK, or external logging service.
| Permission | Why it is needed |
|---|---|
storage |
Save settings, citation cards, memory pins, conversation journals, and license state locally in chrome.storage.local. |
activeTab (added in v3.0.3) |
Reserved for future user-invoked capture flows. Currently a no-op fallback; screenshot capture actually relies on <all_urls> host permission (see below) because captureVisibleTab is triggered from in-page UI buttons rather than the extension icon. |
https://gemini.google.com/* (host permission) |
Inject the reading layer on Gemini pages. The Extension reads visible DOM text only; it does not intercept Gemini's internal APIs. |
https://api.gumroad.com/* (host permission) |
Send the Pro license verification request described in Section 3. |
<all_urls> (host permission) (added in v3.0.11) |
Required by Chrome's captureVisibleTab API — not used to access any website other than Gemini. Chrome's screenshot API mandates either <all_urls> or activeTab, and specific host permissions (like gemini.google.com/*) are not accepted for this API. Because the Extension's content scripts only match https://gemini.google.com/* (see manifest.json → content_scripts.matches), no code runs on any other site and no data from other sites is read, stored, or transmitted. The sole use of this permission is the screenshot feature on the Gemini tab you have already opened. |
The Extension does not set cookies, does not track you across websites, does not fingerprint your device, and does not display advertising.
The Extension is not directed at children under 13. It does not knowingly collect data from children.
chrome.storage.local. You can also
clear individual pieces from the popup (citation cards, memory pins) or
deactivate your Pro license via the popup.All data is stored locally and persists only as long as you keep the Extension installed. Uninstalling the Extension triggers Chrome's standard extension-data cleanup. The Extension also auto-purges temporary handoff data after 60 seconds as a safety measure.
gemini.google.com, which has its own privacy policy governing
how Google handles your conversations.
Material changes will be reflected by bumping the "Last updated" date at the top of this page and, when significant, noted in the release notes of the corresponding Extension version.
Privacy questions or data-deletion requests: open an issue at github.com/WXY395/gemini-reading-assistant/issues.
Gemini Reading Assistant（下稱「本擴充功能」）是在 gemini.google.com 對話頁面上加入導航、搜尋、引用、匯出、 截圖等閱讀輔助功能的 Chrome 擴充。本政策說明本擴充接觸到哪些資料、 儲存在哪裡、以及（若有）哪些資料會離開你的裝置。
簡短版：所有資料都在你本機瀏覽器內，唯一的對外連線 是當你啟用 Pro 授權時向 Gumroad 發送的授權驗證請求。
| 資料 | 儲存位置 | 用途 |
|---|---|---|
| 擴充功能設定（模組開關、Gemini plan 選擇、自訂 prompt） | 本機 chrome.storage.local |
記住你的偏好設定。 |
| 引用卡與記憶錨點（你主動儲存的文字） | 本機 chrome.storage.local |
在 Gemini 中再次使用已保存的段落。 |
| 對話快照與 journal（可見訊息的自動保存） | 本機 chrome.storage.local |
還原上下文、支援 Markdown / TXT / JSON 匯出。 |
| Pro 授權資料（授權碼、驗證時間戳、Gumroad 回傳的購買 email） | 本機 chrome.storage.local |
在本機解鎖 Pro 功能，避免每次都要連線驗證。 |
| 截圖圖片（Ctrl+Shift+X 功能） | 完全不儲存，直接透過瀏覽器原生下載機制下載到你的裝置。 | 讓你把對話片段存成 PNG 檔。 |
本擴充只會發出一種對外網路請求：
https://api.gumroad.com/v2/licenses/verify。Gumroad 會回傳
授權是否有效、以及（可選）購買 email 讓 UI 顯示帳戶持有者。
對話內容永遠不會被傳送。
本擴充不使用任何分析工具、遙測、崩潰回報、廣告 SDK 或外部記錄服務。
| 權限 | 用途 |
|---|---|
storage |
在本機 chrome.storage.local 儲存設定、引用卡、記憶錨點、對話 journal、授權狀態。 |
activeTab （v3.0.3 新增） |
保留供未來「使用者明確激活擴充」流程使用。目前為無作用的 fallback；實際截圖依賴下方的 <all_urls>（因為截圖是從頁內 UI 按鈕觸發，而非點擊擴充 icon）。 |
https://gemini.google.com/* |
在 Gemini 頁面注入閱讀層。本擴充僅讀取可見 DOM 文字，不攔截 Gemini 的內部 API。 |
https://api.gumroad.com/* |
發送第 3 節所述的 Pro 授權驗證請求。 |
<all_urls> （v3.0.11 新增） |
此權限僅為 Chrome captureVisibleTab API 的硬性要求，不用於存取 Gemini 以外的任何網站。 Chrome 的截圖 API 強制要求 <all_urls> 或 activeTab 其中之一，具體 host permission（如 gemini.google.com/*）不被此 API 接受。由於本擴充的 content script 僅註冊於 https://gemini.google.com/*（見 manifest.json 的 content_scripts.matches），不會在任何其他網站執行任何程式碼，亦不會讀取、儲存或傳送其他網站的資料。此權限唯一用途是在你已開啟的 Gemini 分頁上擷取螢幕畫面。 |
本擴充不設定 Cookie、不進行跨站追蹤、不做裝置指紋識別、不顯示任何廣告。
本擴充不以 13 歲以下兒童為對象，亦不會刻意蒐集兒童資料。
chrome.storage.local 的資料。你也可以在 popup 清除個別引用卡、記憶錨點，或取消 Pro 授權。所有資料皆為本機儲存，僅在你仍安裝本擴充功能時存在。解除安裝會觸發 Chrome 標準的 擴充資料清除流程。此外，臨時的 handoff 資料會在 60 秒後自動清除作為安全措施。
gemini.google.com，Google 對你的對話有自己的隱私政策。任何實質變更會反映在本頁頂部的「Last updated」日期；若變更重大，將同步寫入對應擴充版本的 release notes。
隱私相關問題或資料刪除請求，請至 github.com/WXY395/gemini-reading-assistant/issues 開 issue。