透過集合功能整理內容
你可以依據偏好儲存及分類內容。
避免 Google 搜尋顯示受遮蓋的資訊
在網路上發布文件和圖片時,您可能會不小心發布肉眼無法立刻察覺的資訊。特別是,您看不到的資訊或是刻意遮蓋的資訊,可能會以某種文件格式包含在內容中,讓搜尋引擎得以一覽無遺。
由於搜尋引擎會為網路上的公開內容 (包括圖片) 建立索引,因此未完全遮蓋的內容還是有可能讓搜尋引擎找到。如果使用者使用螢幕閱讀器等輔助技術,將會更容易取得這類看似「隱藏」的內容,而常見的圖片辨識技術 (例如光學字元辨識 (OCR)) 同樣也可能讓使用者搜尋到這類內容。
有些人會選擇將文字字型縮小並使用與背景相同的顏色,或者以圖片遮住文字,雖然這類方法能讓使用者無法以肉眼看到資訊,但實際上並未真正遮蓋內容,因此無法避免搜尋引擎為其建立索引或顯示在搜尋結果中。
同樣地,某些文件類型會以各種形式包含我們無法立刻看見的資訊。這些文件類型可能會含有文件的變更記錄,可讓使用者查看經過遮蓋或修改的文字,也可能保留資訊受到裁剪或遮蓋前的完整版圖片。甚至,如果檔案含有中繼資料,雖然使用者無法立刻看見,但其中也可能列出曾經存取或編輯過檔案的使用者名稱。
即使文件是由其他檔案匯出,或曾在不同格式之間轉換,上述這些資訊也全都會保留下來。因此,如果您需要從檔案中移除特定資訊,請務必在檔案公開發布之前完全移除該資訊。
以下提供一些最佳做法,說明如何適當地遮蓋您不想要 Google 建立索引的文件資訊,並讓其他人無法透過 Google 搜尋找到這些資訊。
在嵌入圖片之前先進行編輯並匯出
Google 搜尋會列出系統在網路上找到的圖片,這些圖片可能位在網頁上,或者內嵌於各種格式的文件中。有些使用者在編輯內嵌的圖片時,會直接使用內嵌文件本身的編輯工具,如果系統將這類圖片與該文件分別建立索引,就會導致遮蓋失敗。因此我們建議,請務必先編輯圖片,再將圖片嵌入文件。我們特別建議您採取以下做法:
- 先裁剪掉不想要出現在圖片中的資訊,再將圖片嵌入文件。某些文件編輯工具 (例如文書處理軟體或投影片製作工具) 會保留您放在公開版文件的未裁剪圖片,因此請務必詳閱工具的說明文件。
-
完整移除或遮蓋圖片中所有文字或不想公開的部分,因為 OCR 系統會將圖片上可見的任何文字轉換為可搜尋的文字。
-
移除所有不需要的中繼資料。
完成本文件提供的建議操作後,請匯出更新後的圖片,或者儲存為非向量或平面化圖片檔案格式 (例如 PNG 或 WEBP),這麼做可以防止圖片的這些部分不小心出現在公開文件中。
先編輯或移除不需要的文字,再轉換為公開檔案格式
產生公開版文件之前,請移除任何您不希望在檔案最終版本顯示的文字,並選擇使用不會保留先前變更記錄的公開格式。以下提供幾項更具體的操作提示:
如果 Google 搜尋已為未遮蓋或不當遮蓋的文件建立索引,該如何處理
-
將目前的文件從發布的網站或位置中移除。
-
對已驗證的網站使用移除工具,從 Google 搜尋移除相關文件。如果需要移除多個文件,請使用網址前置字元。如果網站已通過驗證,網址移除程序通常不到一天就會完成。如此一來,當使用者搜尋遮蓋的內容時,就不會看到相關文件。
-
請透過不同網址代管已完整遮蓋的文件。因為重新檢索網址以及更新搜尋索引中的網址需要一些時間,這麼做能確保所有新建立的索引版本依據都是新版文件,而非舊版文件。請記得更新指向這些文件的任何連結。
-
如有任何其他網站也代管這些遮蓋不完全的文件,請與該網站聯絡,要求對方移除這些文件。他們可以透過自己 Search Console 帳戶中的「移除工具」執行移除動作,或者您也可以使用「移除過舊的內容」工具要求 Google 系統更新搜尋結果。
-
等待網址移除要求過期。在系統更新 Google 搜尋索引中的網址後,或大概經過 6 個月時間,要求便會過期。
除非另有註明,否則本頁面中的內容是採用創用 CC 姓名標示 4.0 授權,程式碼範例則為阿帕契 2.0 授權。詳情請參閱《Google Developers 網站政策》。Java 是 Oracle 和/或其關聯企業的註冊商標。
上次更新時間:2025-08-04 (世界標準時間)。
[null,null,["上次更新時間:2025-08-04 (世界標準時間)。"],[[["\u003cp\u003eAvoid unintentionally publishing sensitive information hidden within documents or images, as search engines can index this data.\u003c/p\u003e\n"],["\u003cp\u003eEmploy proper redaction techniques like cropping, obscuring, or removing sensitive content before publishing files, rather than relying on visual concealment methods.\u003c/p\u003e\n"],["\u003cp\u003eEdit images and remove unwanted text before embedding them into documents to prevent unintended exposure of redacted content.\u003c/p\u003e\n"],["\u003cp\u003eUse the Google Search Console's removal tools if improperly redacted documents are indexed to quickly remove them from search results.\u003c/p\u003e\n"]]],["To keep information out of Google Search, properly redact documents before publishing. Edit images before embedding them, cropping unwanted parts, removing text, and deleting metadata. Remove unwanted text from documents, using redaction tools instead of merely covering text. Export images in non-vector formats. For indexed, improperly redacted content, remove the live document, use Google Search Console's Removals tool, host a properly redacted version under a new URL, and contact other hosts to remove the documents.\n"],null,["# Keep Redacted Information out of Google | Google Search Central\n\nKeep redacted information out of Google Search\n==============================================\n\n\nWhen publishing documents and images on the web, you may unintentionally publish information\nbeyond what is immediately visible to the human eye. In particular, information that you might\nnot see, or that was intended to be redacted, might be included in some document formats and\nvisible to search engines.\n\nBecause search engines index public material on the web, including images, content that is\nnot completely redacted can potentially be findable in search engines. Assistive technologies\nlike screen readers can make this seemingly \"hidden\" content more easily accessible, and\ncommon image understanding techniques like optical character recognition (OCR) similarly make\nit possible to search for this content.\n\nEven though putting text in a tiny font, using a font color that's the same as the background\nthe text is on, or covering text with an image may make something invisible to the human eye,\nthese methods don't actually redact material in a way that prevents search engines from\nindexing it and making it findable.\n\n\nSimilarly, some document types include information in various ways that aren't immediately\nvisible. They might include the document's change history, allowing users to see text that has\nbeen redacted or altered. They might retain the full versions of images that contain cropped\nor redacted information. There might also be metadata that's included in a file, which is not\nimmediately visible, that may list the names of people who accessed or edited the file.\n\n\nAll of this information can remain even when a document is exported or converted from one\nformat to another. If you need to remove information from a file, it's critical that the\ninformation is removed completely from the file before that file is made public.\n\n\nHere are some best practices for how to appropriately redact information from documents that\nyou don't want to be indexed and made discoverable via Google Search.\n\nEdit and export images before embedding them\n--------------------------------------------\n\n\nGoogle Search lists images that it finds across the web, both those that are on web pages or\nthose that are embedded into various document formats. Embedded images are sometimes edited\nusing only the containing document's editing tools. This can cause this redaction to fail when\nan image is indexed apart from the document. That is why it's best to edit images before\nembedding them into a document, not after. In particular:\n\n- Crop out unwanted information from images before embedding them into documents. Some document editing tools (such as word processors or slide creation tools) will maintain any uncropped images that you use in the public version of the document, so be sure to review the tool's documentation thoroughly.\n- Completely remove or obscure any text or other non-public parts of the image, as OCR systems may turn any image text seen into searchable text.\n- Remove any undesired metadata.\n\n\nAfter following the suggestions in this document, export or save the updated images as non-vector or\nflattened image file formats such as PNG or WEBP. This prevents those parts of the images from\nbeing inadvertently included in a public document.\n\nEdit or remove unwanted text before moving to a public file format\n------------------------------------------------------------------\n\n\nBefore you generate the public document, remove any text that you don't want displayed in the\nfinal version of the file. Move to a public format that does not keep your previous change\nhistory. Here are more specific tips:\n\n- Use proper document redacting tools if a file needs to have information redacted. For example, avoid placing black rectangles over text as a redaction method, as this can result in the text still being included in the public document.\n- Double-check the document metadata in the public file.\n- Follow the [document redaction best practices](https://www.google.com/search?q=document+redaction+best+practices) for the format that you are using (PDF, image, etc).\n- Consider information in the URL or file name itself. Even if a part of a website is [blocked by robots.txt](/search/docs/crawling-indexing/robots/intro), the URLs may be indexed in search (without their content). Use hashes in URL parameters instead of email addresses or names.\n- Consider using authentication to limit access to the redacted content. Serve the resulting login page with a [`noindex` robots `meta` tag](/search/docs/crawling-indexing/block-indexing) to block indexing.\n- When publishing, make sure that the website is [verified in Google Search Console](https://support.google.com/webmasters/answer/9008080). This allows quick removal action, if needed.\n\nWhat to do if unredacted or improperly redacted documents are indexed in Search\n-------------------------------------------------------------------------------\n\n1. Remove the live document from the website or location where you published it.\n2. Use the [Removals tool](https://support.google.com/webmasters/answer/9689846) for the verified site to remove the documents in question from Search. Use a URL prefix if you need to remove many documents. For verified sites, a URL removal generally takes less than a day. This prevents the document in question from appearing for any searches for redacted content.\n3. Host the properly redacted document under a different URL. This makes sure that any newly indexed version is of the new document, and not an older version of the document (since recrawling of URLs and updating them in a search index can take a bit of time). Update any links to those documents.\n4. Contact any other site that may also be hosting the improperly redacted documents and ask them to take them down as well. Ask them to use the Removals tool in their Search Console account, or you can use the [Outdated Content tool](https://support.google.com/webmasters/answer/7041154) to ask Google's systems to update the search results.\n5. Allow the URL removal requests to expire (this happens after the URLs were either updated in the Google Search index, or after about 6 months)."]]