How Local Browser Tools Protect Your Files
A privacy-first collection of browser-based productivity utilities.
What “processed locally” means
When you choose a file, JavaScript running in the page reads the file through browser APIs. The tool then transforms the data in memory and creates a new downloadable file.
Why it improves privacy
There is no need to send the document to an application server simply to count words, resize an image, convert CSV, or generate a password. This reduces unnecessary data transfer.
What still uses the network
The main site files need to be downloaded when you open the website. PDF and QR pages also load an open-source library from a public CDN unless those libraries are later bundled into the site package.
Good practice
- Use an updated browser.
- Keep backups of important originals.
- Review output files before deleting originals.
- Avoid processing highly sensitive files on shared computers.
How to verify it yourself
You do not have to take the claim on trust. Open your browser's developer tools, switch to the network panel, then load a file into any tool and process it. You will see the page's own assets and, on some pages, one library fetched from a public CDN — and no request carrying your file. Disconnecting from the network after the page has loaded is an even blunter test: most tools keep working, because there is nothing for them to call.
Which browser capabilities do the work
FileReader and the File API read the file you selected. Canvas decodes, redraws and re-encodes images. Web Crypto computes SHA checksums using the same implementation the browser uses for TLS. Local Storage holds your language choice and the contents of the notes and to-do tools. None of these APIs transmits anything on its own; each operates on data already in the page.
The limits of local processing
Everything happens in one browser tab, so memory is the ceiling and very large batches can exhaust it. Formats the browser cannot decode natively, such as camera RAW, will not open. And local processing protects the transfer, not the device: on a shared or managed computer, the file you downloaded is still on that machine's disk, and screen recording or endpoint software sees what you see.
What this does and does not promise
It means your documents are not sent to us, not stored by us, and not available to us — there is no server that could hold them. It is not a claim about the security of your own device, your network, or any file you subsequently upload somewhere else. For genuinely sensitive material, the sensible habit is the same as always: work on a machine you control, and check what a document still carries — the metadata viewers here will tell you — before you pass it on.
「本地處理」是什麼意思
當你選擇檔案時，頁面中的 JavaScript 會透過瀏覽器 API 讀取檔案。工具會在記憶體中轉換資料，再建立可下載的新檔案。
為什麼能改善隱私
字數統計、圖片尺寸調整、CSV 轉換或密碼產生等工作，不需要把文件傳送到應用程式伺服器，可減少不必要的資料傳輸。
哪些部分仍會使用網路
開啟網站時需要下載主要網站檔案。PDF 與 QR Code 頁面也會從公共 CDN 載入開源程式庫，除非日後將這些程式庫直接打包進網站。
使用建議
- 使用新版瀏覽器。
- 保留重要原始檔案的備份。
- 刪除原檔前先檢查輸出結果。
- 避免在共用電腦處理高度敏感檔案。
如何自己驗證
你不必只憑信任。打開瀏覽器的開發者工具、切到網路面板，載入檔案並執行任一工具，你會看到頁面本身的資源，以及部分頁面從公開 CDN 取得的一支函式庫——但沒有任何請求帶著你的檔案。更粗暴的測試是：頁面載入後直接斷網，多數工具照樣運作，因為它們本來就沒有要呼叫的對象。
實際做事的瀏覽器能力
FileReader 與 File API 負責讀取你選的檔案；Canvas 負責解碼、重繪與重新編碼圖片；Web Crypto 計算 SHA 校驗值，用的是瀏覽器處理 TLS 時的同一套實作；Local Storage 則保存語言選擇與記事本、待辦清單的內容。這些 API 本身都不會主動傳送任何資料，它們處理的都是已經在頁面裡的內容。
本機處理的限制
所有運算都在同一個瀏覽器分頁中，因此記憶體就是上限，超大批次可能會撐爆它。瀏覽器原生無法解碼的格式（例如相機 RAW）也開不起來。此外，本機處理保護的是「傳輸」而非「裝置」：在共用或受管控的電腦上，你下載的檔案仍留在那台機器的硬碟裡，螢幕錄製或端點管理軟體看得到你看到的一切。
它承諾什麼、不承諾什麼
它代表你的文件不會傳送給我們、不會被我們儲存，我們也拿不到——因為根本沒有能存放它們的伺服器。但它並不是對你自己的裝置、你的網路，或你之後上傳到別處的檔案所作的安全保證。真正敏感的資料，好習慣始終如一：在自己能掌控的機器上處理，並在轉交之前先確認文件裡還帶著什麼——本站的中繼資料檢視工具就能告訴你。