Detect and remove comments, highlights, and drawings — while links and forms stay protected by default. Runs entirely in your browser.
Drop your PDF to scan for annotations
Click to browse or drag & drop
🔒 Your files never leave your device — scanning and removal both run locally in your browser.
Password protected PDF
Enter the password to open this PDF.
5 attempt(s) remaining
Analyzing your PDF…
Loading PDF engine
No removable annotations found
This PDF has no comments, highlights, drawings, stamps, or similar annotation objects. If you expected to see marks that aren't listed, they may be baked into a scanned page's image rather than stored as removable annotations — see the FAQ below.
Recommended: Remove Comments & Markups
Removes sticky notes, comments, highlights, underlines, strikeouts, and drawings from every page. Stamps, links, form fields, and attachments are left untouched.
Select categories to remove. Links and Form Fields require explicit confirmation because removing them can affect document functionality.
| Page | Type | Author | Content | Modified |
|---|
No annotations selected for removal
You'll download the cleaned PDF on the next screen.
Annotations removed successfully
—
TL;DR
Four steps. No install, no account. Files never leave your browser.
Drag your file into the box above, or click to browse. Nothing uploads to a server — the file is read directly by your browser.
Every annotation is detected and classified automatically — see counts by type, which pages they're on, and full details in the inspector.
Click the recommended one-click action, or open Advanced to pick exact categories, individual annotations, and a page range.
The cleaned PDF is validated — reopened, page count checked, and re-scanned — before your download link is enabled.
Definition
In the PDF specification, an annotation is a distinct object placed on top of a page's content — separate from the page's actual text, images, and vector graphics. A highlight, a sticky note, a hand-drawn circle, a stamp, a hyperlink, and a fillable form field are all technically "annotations," even though they look and behave very differently to a reader. Ordinary page content is never an annotation: normal body text, embedded photos, background graphics, and page backgrounds are part of the page itself and are never touched by this tool.
The PDF spec defines around two dozen annotation subtypes. PDFcrest groups them into plain-language categories so you don't need to know the spec to use the tool correctly:
Why PDFcrest
Every removal edits the PDF's real annotation list — nothing is hidden with a filter or rasterized away.
Every detected object is identified by its real PDF subtype and grouped into eight plain-language categories — not a single "remove everything" bucket.
The one-click safe action never removes hyperlinks or interactive form fields. Including them requires an explicit, separate confirmation.
Signature form fields are detected and permanently excluded from every removal mode, with a clear warning about how any PDF modification can affect an existing digital signature.
Uncheck any single annotation in the inspector, or scope the whole cleanup to a custom page range like 1-5,8,10-12 — everything outside your selection is left completely untouched.
Scanning and removal both run in your browser tab using a WebAssembly PDF engine. Your file's content is never transmitted anywhere.
The generated PDF is reopened, its page count checked, its modified pages re-scanned to confirm removal, and a representative page rendered — before the download link is enabled.
Is This For You?
Anyone handing a PDF off to someone outside the review — editors finalizing a manuscript, lawyers clearing redline comments before a document leaves the firm, teachers stripping margin notes from a shared reading, or anyone who received a heavily-annotated file and just wants a clean copy back.
Right after internal review wraps up and right before the file goes to its actual audience — a client, a student, the public record — so leftover sticky notes, highlights, and stamps don't ship by accident. Not while a document is still mid-review, since that would erase comments people are actively discussing.
Skip it if what you're trying to erase isn't a real annotation object to begin with — a highlighter stroke or handwritten note on a scanned page is baked into that page's image pixels, not a PDF annotation, so no annotation remover can touch it. That needs image editing or a fresh scan instead.
Comparison
A PDF reader's "hide comments" or "hide markup" view setting only changes what's displayed on your screen in that app — the annotations are still stored in the file, and reappear the moment the file is opened elsewhere or that setting is turned off. PDFcrest instead deletes the annotation objects from the PDF's actual structure, so the result is genuinely annotation-free in any viewer, on any device.
| Feature | PDFcrest | PDF reader "hide markup" view | Typical online tool |
|---|---|---|---|
| Price | Free, unlimited | Free, viewer-only | Free tier capped |
| Actually deletes annotations from the file | Yes | No — only hides them from view | Sometimes |
| Classifies annotations by type | 8 categories, per-type counts | N/A | Rarely more than "all" |
| Protects links & forms by default | Yes, with explicit opt-in to include them | N/A | Often removes everything |
| Per-annotation inspector | Search, filter, individually exclude | Basic list at best | Rarely available |
| Signature field protection | Detected and always excluded | N/A | Rarely addressed |
| File privacy | 100% browser-based | Local app | Server upload typical |
| Output validated after processing | Reopened, re-scanned, page rendered | N/A | Rarely verified |
Use Cases
Strip a document of every reviewer's markup before sending a clean final version, while keeping any embedded links and forms intact.
Remove privileged review comments and redline markup from a contract or filing before it's shared externally.
Clear highlights and margin notes from a shared reading before redistributing it, or before returning a graded assignment's clean copy.
Finalize a proposal or report by removing internal comment threads while preserving the document's links and any fillable sections.
Remove reviewer annotations and, where appropriate, printer/system marks before a document goes to print.
Produce a clean archival copy of a document stripped of ad-hoc review markup, without disturbing its underlying text or structure.
Protected By Default
Not every annotation is a "comment" someone left by mistake. A Link annotation is often essential document navigation — a table of contents entry, a citation, a reference to another page. A Widget annotation is a fillable form field — a text box, checkbox, or dropdown that makes a PDF interactive. Deleting either by default, the way some "remove everything" tools do, can silently break a document's actual functionality.
PDFcrest's recommended cleanup action never touches links or form fields. To remove either, you have to open Advanced options and explicitly confirm — a deliberate extra step designed to make sure it's a decision, not an accident.
Signature fields are a further special case. A Widget whose field type is "signature" is detected and permanently excluded from removal in every mode, including the Advanced "Form Fields" category — there's no way to select a signature field for deletion in this tool. That said, digitally signing a PDF cryptographically covers a specific byte range of the original file. Saving any modified copy of a signed PDF — even one where only unrelated comments were removed — can invalidate that signature, because the file's bytes no longer match what was signed. PDFcrest surfaces a clear warning whenever a signature field is detected so you know before you download.
Important Limitation
Only if the mark actually exists as an annotation object. Two common situations look similar on-screen but are technically very different:
PDFcrest is honest about this distinction rather than promising a result it can't deliver: annotation removal targets genuine, structured annotation objects. It's not a substitute for image editing or OCR-based content redaction.
What Stays Intact
Because annotations are removed as discrete objects from each page's annotation list — not by re-rendering, flattening, or rasterizing the page — the underlying page content is left alone. Text stays selectable, fonts and embedded images are untouched, vector graphics are unaffected, and page dimensions, orientation, and page count don't change. Any invisible OCR text layer on a scanned page is ordinary page content, not an annotation, so it's preserved exactly as it was.
What annotation removal is not the same as: it doesn't strip document metadata (title, author, subject), doesn't remove embedded files that aren't attachment annotations, doesn't remove document-level JavaScript, and isn't a general PDF sanitization or redaction tool. If you need to clear metadata specifically, PDFcrest's Edit PDF Metadata tool handles that separately.
FAQ
Glossary
| Term | Definition |
|---|---|
| Annotation | A distinct PDF object placed on a page — a comment, markup, drawing, stamp, link, or form field — separate from the page's underlying text and graphics. |
| Markup annotation | An annotation that visually marks up existing content, such as a Highlight, Underline, StrikeOut, or Squiggly. |
| Popup | A small annotation window holding a comment's text, author, and date — linked to its parent annotation and removed alongside it. |
| Widget | The annotation subtype used for interactive form fields: text boxes, checkboxes, radio buttons, dropdowns, buttons, and signature fields. |
| AcroForm | The document-level structure that ties together all of a PDF's form field widgets. |
| Appearance stream | The saved visual rendering of an annotation, used so viewers can display it without recomputing its appearance. |
| Flattened annotation | An annotation that has been permanently merged into a page's regular content, after which it no longer exists as a separate, removable object. |
| Digital signature | A cryptographic signature covering a specific byte range of a PDF file; any later modification to the file can invalidate it. |
| Object graph | The network of linked PDF objects (pages, annotations, fonts, images) that make up a document's internal structure. |
| Client-side processing | Computation that happens on the user's own device, in the browser, rather than being sent to a server. |
References & Further Reading
Last updated: