你的 PDF 可能没真正涂干净——一招自查
很多人给 PDF「打码」的方式，是在敏感文字上画一个黑色矩形，然后导出。看起来涂掉了，其实那段文字还原封不动地待在文件里。
30 秒自查
- 打开你已经「涂黑」的 PDF；
- 按 ⌘A / Ctrl+A 全选，再 ⌘C / Ctrl+C 复制；
- 粘贴到任意文本编辑器里。
如果被涂黑的内容出现在粘贴结果里——你的脱敏是假的。任何拿到这份文件的人都能用同样的方式还原它，甚至直接用「搜索」就能命中被盖住的词。
为什么会这样
PDF 页面由一层层对象组成：文字对象、图片对象、图形对象。画黑条只是在最上面又叠了一个黑色矩形，底下的文字对象一个字都没少。复制、搜索、无障碍朗读，读到的都是底层文字，而不是你看到的黑条。
此外文件的元数据（标题、作者、备注）也常常藏着不想公开的信息，单纯画黑条根本管不到它们。
真实事故一直在发生
2019 年，Paul Manafort 的律师向法院提交了一份「已脱敏」的文件，记者把黑条下的文字直接复制了出来，多段关键内容因此曝光。2020 年，Ghislaine Maxwell 的证词文件同样被发现黑条可选中复制。学术界 2022 年的一项研究扫描了数十万份公开 PDF，发现数千份文件的脱敏可以被还原。这类新闻每隔几个月就会再来一次。
真脱敏应该做到什么
- 删除，而不是遮盖：被涂黑区域的底层文字与图像要从文件里彻底移除；
- 清除元数据：标题、作者等信息一并清空；
- 可验证：处理完能证明「已无可提取内容」，而不是让你凭感觉相信。
用 Blackbar 免费做真脱敏
Blackbar 的脱敏工具把被涂黑的页面整页重建：底层文字与图像被真正删除，元数据被清除，导出后立刻重新解析文件、当场给你一份验证报告（可提取文字块 = 0）。
更重要的是：你的文件从头到尾不会离开你的设备。这个站点是纯静态页面，没有服务器可以接收上传——要涂黑的本来就是敏感文件，凭什么先传给别人？断网也能用（可安装为应用）。
30 秒检测我的 PDF →
直接去脱敏 →
Your “redacted” PDF probably isn’t — a 30-second self-check
The way most people redact a PDF is to draw a black rectangle over the sensitive text and export. It looks redacted. The text is still sitting in the file, untouched.
The 30-second check
- Open your “redacted” PDF;
- Press ⌘A / Ctrl+A to select all, then ⌘C / Ctrl+C;
- Paste into any text editor.
If the blacked-out content shows up in the paste — your redaction is fake. Anyone with the file can do exactly the same, or simply hit the covered words with search.
Why this happens
A PDF page is a stack of objects: text objects, images, graphics. Drawing a black bar just adds one more rectangle on top. The text objects underneath are all still there. Copy, search and screen readers all read the underlying text, not the pixels you see.
Document metadata (title, author, comments) often leaks too — a black bar can’t touch it.
This keeps happening for real
In 2019, Paul Manafort’s lawyers filed a “redacted” court document; journalists copy-pasted the text right out from under the black bars. In 2020, the Ghislaine Maxwell deposition had the same flaw. A 2022 academic study scanned hundreds of thousands of public PDFs and found thousands of recoverable redactions. This story returns to the news every few months.
What real redaction requires
- Deletion, not covering: the text and images under the box must be removed from the file;
- Metadata scrubbing: title, author and friends wiped;
- Proof: after processing, show that nothing extractable remains — don’t ask for faith.
Do it for free with Blackbar
Blackbar’s redaction tool rebuilds every redacted page so the underlying text and images are truly gone, scrubs the metadata, then re-parses the exported file and hands you a verification report (extractable text items = 0).
And crucially: your file never leaves your device. This site is static — there is no server that could receive an upload. The file you’re redacting is sensitive by definition; why would you send it to someone else first? Works offline too (installable as an app).
Check my PDF in 30 seconds →
Go redact →