Clean PDF
See everything a PDF carries before you share it, then remove it: properties, attached files, scripts, hidden layers and earlier versions.
or drop files here
You can also paste with Ctrl+V
A document we wrote for this: nothing of yours is used.
Processed on your device. Files are never uploaded.
Quick facts
- Takes
- One PDF, or several with Pro. A password-protected PDF is refused until it is unlocked.
- Gives
- The same pages without what traveled unseen: properties and XMP metadata, attached files, scripts and the actions that start programs or send data, the content of layers hidden by default, stored thumbnails and programs’ private data, and every earlier version saved in the file. Comments, form fields and bookmarks go too when you choose.
- Runs
- In your browser. The file is never uploaded, and nothing is kept anywhere.
- Plan
- Free, with no account and no watermark: 10 tasks a day, on files up to 50 MB and 100 pages. Pro has no daily count and takes files up to 500 MB, with no page limit.
- Says
- What it removed, counted: how many properties, files, scripts, layers and earlier versions. Then it reads the clean file again, with the same reader as See what the file carries, and says what the file still holds: what you chose to keep, and anything it was asked to remove and could not. It also says what it never looks at: what the pages show.
What goes, and what stays
The pages stay as they are. What the document keeps beside them goes: the information dictionary and the XMP packet, the files attached to it and to its comments, its named scripts and the action it runs on opening, and every link, field or page action that runs a script, starts a program, opens another file or sends data somewhere. Links to web pages and to other pages of the document keep working.
A layer the document hides by default loses what it draws, and then the layer itself goes, so no reader’s layers panel can bring it back and its name no longer tells what it held. The tool takes out only what such a layer paints, the text, pictures and shapes, and keeps every change it makes to the drawing state, so everything drawn after it looks exactly as before. Layers shown by default stay.
The clean file is written afresh, which always drops the earlier versions an incremental save leaves inside a file. Cleaning changes the file, so a digital signature no longer matches it: sign the clean copy instead. What the pages themselves show is content, not hidden data; remove it with Redact PDF first, then clean. To see what a file holds before or after, open it in Inspect PDF.
Measured on our sample file
| Sample | Input | Output | Removed | Time |
|---|---|---|---|---|
| Offer letter, draft | 29 kB | 22 kB | 7 properties, the XMP metadata, 1 attached file, 1 script or action, 1 hidden layer, 1 earlier version, 1 thumbnail, 1 item of private data | 0.1 s |
What the file carries, first
Choose See what the file carries above the file for the full report before anything is removed: properties, attachments, scripts, layers, fonts, earlier versions, and where each is.
Quick facts
- Takes
- One PDF, up to 500 MB. A password-protected PDF is read once its password is typed on the file card.
- Gives
- A report of what the file carries: its version, pages and sizes, protection and PDF/A claim; its properties and XMP metadata; attached files, scripts, layers and earlier versions saved in the file; and its fonts, pictures, bookmarks, form fields and annotations. The report downloads as JSON.
- Runs
- In your browser. The file is never uploaded, and nothing is changed.
- Plan
- Free, with no account, and never counted: opening a PDF to read it is not a task.
Where a PDF keeps what it does not show
A PDF is more than its pages. Beside them it keeps an information dictionary (title, author, the program that made it) and often an XMP packet that repeats it and adds more, such as everyone who worked on the file. It can carry files attached to the document or to a comment, scripts that run when it opens or when a link is pressed, layers that the document hides unless someone switches them on, and stored thumbnails. Programs that save a PDF incrementally add each change to the end of the file and leave what was there before behind it, where the right tool reads the earlier version back.
The tool reads each of these places and lists what it finds; a summary at the top counts the kinds of hidden data. It runs nothing it finds: a script is reported where it is, not executed. To take it all out, choose Clean this file, which opens Clean PDF with the same file. What the pages themselves say, a name in a footer for one, is content: that is for Redact PDF.
Measured on our sample file
| Sample | Kinds of hidden data found | Time to the report |
|---|---|---|
| Offer letter, draft | 8 | 1.2 s |
What it cannot do
- Remove what the pages show. A name in a header or a figure in a table is content; redact it first with Redact PDF.
- Keep a digital signature valid. A cleaned file is a changed file; sign it again afterwards.
- Remove content from layers that are shown, or from layers whose visibility follows a formula. Only layers hidden by default lose what they draw.
- Clean a password-protected file without its password. Unlock it first, then protect the clean copy again if you need to.
- Find text hidden on the pages themselves, such as white text on white or text under a picture. It reports what is beside the pages; the pages are what the viewer shows.
- Tell whether a script is harmful. It says where scripts and risky actions are; it does not run them or judge them.
- Read a password-protected file without its password. Type the password on the file card first.
If something goes wrong
- The original file was digitally signed. Signatures don't carry over to a changed file.
- That is expected: a signature certifies one exact file. Clean first, then sign the clean copy. Open Sign PDF with a certificate
- It says there was nothing to remove
- The file had none of what you chose. It has still been written afresh, which drops any earlier versions; Inspect PDF shows what it holds. Open Clean PDF
- This PDF is protected. Enter its password to continue.
- Open it in Unlock PDF with its password, clean the unlocked copy, and protect it again if needed. Open Unlock PDF
- This PDF is protected. Enter its password to continue.
- Type the password on the file card; it never leaves your device.
- The report lists properties I never set
- Programs write them on their own: the producer, the dates, sometimes the name of the account that made the file. Clean PDF removes them all. Open Clean PDF
- This PDF is damaged and can't be read.
- Run it through Repair PDF first; the repaired copy reads normally. Open Repair PDF
About Clean PDF
Clean PDF removes what a PDF carries without showing it: properties and metadata, attached files, scripts, the content of hidden layers, stored thumbnails, editing programs' private data and the earlier versions saved inside the file. Comments, form fields and bookmarks go too if you choose. The pages stay as they are, and it all happens in your browser.
How to use Clean PDF
Add the PDF; nothing is uploaded.
Choose what to remove; the usual hidden data is chosen for you.
Choose Clean PDF, read what was removed, and download the clean file.
Frequently asked questions
- Does cleaning change the pages?
- No. What the pages show stays as it is; only what a layer hidden by default draws is removed, since nobody sees it anyway.
- Is cleaning the same as redacting?
- No. Redact PDF removes what is on the pages, under the boxes you draw; Clean PDF removes what is beside them. Before sharing a sensitive document, redact the pages first, then clean the file.
- What happens to the earlier versions?
- They always go: the clean file is written afresh, with only the current version in it.
- Will a digital signature still be valid?
- No. Cleaning changes the file, so a signature no longer matches it, and the result says so. Sign the clean file instead.
- Can I clean many files at once?
- Yes, with Pro: add several files and each is cleaned with the same choices, the results in one ZIP.
- What counts as hidden data?
- Anything a reader does not show on the page: the properties and XMP metadata, attached files, scripts, layers hidden by default, comments, form fields, stored thumbnails, editing programs' private data, and earlier versions of the document saved in the same file.
- What is an earlier version?
- Programs that save a PDF incrementally add each change to the end of the file, and what was there before stays behind it, where the right tool reads it back. Clean PDF writes the file afresh, so only the current version remains.
- Does it tell me whether the file is safe?
- It shows where scripts and actions that start programs or send data are, so you know they are there. It is not a virus scanner, and it runs nothing.
Related tools
- Redact PDFRemove sensitive text and pictures for good: find emails, phone, card, Aadhaar, PAN and IBAN numbers at once, or draw boxes.
- Edit PDF propertiesChange or remove the title, author, keywords and other details stored in a PDF.
- Protect PDFAdd a password with strong AES-256 encryption.
- Compress PDFMake a PDF smaller, down to an exact size if you need one.