How to Replace Text in a PDF

To replace text in a PDF for free, cover the old words with a background-colored rectangle and retype the replacement in a matching font - the visual swap takes minutes in a browser-based editor. Replacing the same term dozens of times? Convert the PDF to Word for true find-and-replace, then convert back. This guide covers both methods, the font-matching checklist that makes edits undetectable, and when sensitive text needs redaction first.

Quick answer: for a handful of replacements, open the free Edit PDF tool, cover each old passage with a background-matched rectangle, and type the new text in a matching font. For replacements across a whole document, use the free PDF to Word round trip and find-and-replace there.

On this page

• Why PDF text replacement is different
• Method 1: Cover and retype (free, browser)
• Method 2: Find-and-replace via Word (free)
• Sensitive text: redact first, then replace
• Replacing text in scanned PDFs
• The matching checklist: font, size, color, position
• Choosing between the methods
• Troubleshooting
• FAQ

Why PDF text replacement is different

In a word processor, a document is sentences: "replace every 2025 with 2026" is one command because the text exists as searchable, structured strings. In a PDF, the page is a drawing: characters are placed glyphs, and a single visible sentence is often stored as several internal runs - sometimes split mid-word to satisfy line-breaking and kerning. There is no sentence object to search coherently, and no reflow engine to accommodate longer or shorter replacements.

Every text replacement in a PDF is therefore one of two operations:

1. Visual replacement - hide the old glyphs and draw new ones. The document's appearance changes; the underlying file structure keeps both texts (the old one hidden). This is cover-and-retype, the free workhorse.
2. Structural replacement - rebuild the document so the text layer itself changes. This is what desktop editors' text-editing modes do (with font-matching constraints), and what the Word conversion does wholesale.

Choosing between them is mostly a function of scale and sensitivity, which is what the methods section operationalizes.

Method 1: Cover and retype (free, browser)

The 80% method: fast, free, private, and visually perfect when done with the matching checklist below.

Step 1. Open the PDF in the Edit PDF tool - local processing, no upload.

Step 2. Cover each old passage. With the rectangle tool, draw a box filled with the page background color (white on most pages; sample tinted backgrounds) directly over the text to replace. Zoom to 200-400% for this: the cover must hug the old text - one pixel of visible old glyph edge reads as an error at arm's length.

Step 3. Type the replacement. Text tool, font/size/color set before clicking, positioned where the old text began. Type the new content, then nudge the box until its baseline aligns with neighboring lines.

Step 4. Verify at 100% zoom, then at 400%. The 100% check is "does it read naturally"; the 400% check is "does anything peek out from under the cover."

Step 5. Download. Your replacement text is real, selectable text; the old text is hidden beneath.

Honest limits: the old text survives in the file (matters for sensitive content - see redaction below), and each replacement is manual. For a wrong price on an invoice or a renamed product on a flyer, this is minutes of work. For thirty instances of a changed term, it is the wrong tool.

Method 2: Find-and-replace via Word (free)

When replacements are numerous or document-wide, the conversion route turns a PDF problem into a Word one - where it was always trivial.

Step 1. Convert the PDF to DOCX with the free PDF to Word tool. The document arrives as flowing text with paragraphs and headings reconstructed.

Step 2. Find and replace natively. Ctrl+H (Word, LibreOffice) or the Docs equivalent: replace every instance of the old term in seconds, with optional case matching and whole-word constraints. This is the step that makes the whole detour worthwhile - and it catches instances your eyes would have missed.

Step 3. Review the reflow. Replacements that change text length shift line breaks; scan the pages once, fix table stragglers, confirm nothing overflows.

Step 4. Export back with Word to PDF. The new file has a genuinely replaced text layer - search, copy and extraction all reflect the new content, with no hidden originals.

Limits, plainly: the rebuild is not pixel-identical to the original (fine for working documents; wrong for signed or certified files), and forms/signatures do not survive the trip. Where the document's integrity matters, Method 1 on a copy is the route - or the two combine: Word route for the working copy, original untouched as the record.

Sensitive text: redact first, then replace

When the text being replaced is sensitive - account numbers, national IDs, client names in a case file - the hiding order matters enormously, and getting it wrong leaks data.

The trap: cover-and-retype hides the old text but keeps it in the file. Anyone with the PDF can select-under-the-box, search, or extract it. For a wrong price that is harmless; for a bank account number it is a data leak waiting for one curious recipient.

The safe sequence:

1. Redact the original. Mark the sensitive text for redaction with a true redaction tool - redaction deletes the underlying text object, not just its appearance. PDFEditorFree's Edit PDF tool includes redaction, and the Sanitize PDF tool strips hidden text and metadata document-wide.
2. Verify the removal. In the saved file, search for the sensitive string: zero results is the pass condition. This ten-second check is the entire difference between redaction and theater.
3. Type the replacement (if one is wanted) on the redacted area, per Method 1's matching checklist.

The same logic applies at document scale: before sharing a PDF that once contained sensitive data anywhere - even on removed pages - run a sanitize pass, because copy-paste history, annotations and form data are all hiding places.

Replacing text in scanned PDFs

Scans have no text objects - the page is a photograph - so "replacement" is purely visual: hide the old pixels, draw the new text.

The workflow: open the scan in the Edit PDF tool, cover the old text with a rectangle matched to the scan's background (scanned "white" paper is rarely pure white - sample or use an off-white like F7F5F0, or the patch will glow), and type the replacement on top. Match the scan's ink: scanned text is usually soft dark gray rather than black, so a pure-black replacement reads sharper than its neighbors - darken slightly instead (333333 is a good default).

The precision trick for scans: zoom to 200% and let the replacement text's baseline sit on the scan's line if the content is handwritten-form style; for body text, match the surrounding line positions rather than exact pixels.

When the scan needs many replacements, OCR changes the economics: recognize the document (OCR tool), convert to Word, replace properly, and rebuild - the rebuild will not match the scan's look, but for working purposes it beats thirty manual patches. Choose by destination: visual fidelity (share the patched scan) versus content utility (OCR route).

The matching checklist: font, size, color, position

An undetectable replacement matches four properties in order of noticeability. Work the list top-down on every replacement:

1. Font category. Sans replaces sans (Helvetica/Arial family), serif replaces serif (Times/Georgia family). A category mismatch is visible across the room. If the original font is identifiable (document properties list embedded fonts), pick its standard twin.
2. Visual size. Compare x-heights, not point numbers: type the replacement, then step size in half-point increments until letter heights match the neighbors optically.
3. Color. Sample or approximate the existing text. Near-black (1a1a1a to 333333) is the common truth behind "black" documents; pure black additions on soft-black originals are one of the two most common tells.
4. Position. Baseline alignment with the line above/below - nudge vertically at 200% zoom until the replacement's feet sit on the same plane. Left edges align to the paragraph's existing margin.

Run the checklist on the two most common failures first (color and baseline) when a replacement "looks off" - it usually resolves in one adjustment.

Choosing between the methods

The two-axis decision behind the table: how many edits (manual scaling limit), and does the hidden-original property matter (sensitivity). Every real job lands in a row.

Troubleshooting

Old text peeks out from under my cover. The rectangle is smaller than the text's bounding box - enlarge it a pixel on each side at 400% zoom, or the font renders wider than it looks (cover letter-spaced runs generously).

Replacement sits a hair above/below the line. Baseline misalignment - select the text box and nudge vertically in single steps at 200% zoom until it sits in the plane of its neighbors.

The cover is visible as a gray box on print. The fill is not 100% opacity or the background color is slightly off. Set opacity to 100% and sample the true background (scanned pages are off-white).

Find-and-replace in Word missed instances. The PDF's text runs split words internally; the conversion usually stitches them, but hyphenated line breaks ("-con tract") evade matching. Run a second pass searching for fragments.

The replaced document's search still finds the old term. Expected for Method 1 - the hidden original is searchable. If that is unacceptable, the fix is redaction (Method 3), not a better cover.

My replacement changed the font after saving. The viewer substituted a missing font. Use the standard substitutes from the matching checklist, which are always available.

Typed text wraps oddly in a narrow column. Position text boxes do not reflow like paragraphs. Break the replacement into multiple boxes aligned line-by-line, or shorten the wording.

Replace-with-care: documents where edits draw scrutiny

Some documents live in worlds where every change is compared against a reference - and replacements there need an audit trail, not just good typography.

Invoices and billing documents. A replaced amount on an invoice you did not issue is a dispute waiting to happen - the issuer's system holds the original, and mismatches flag your copy. The professional path for corrections is a credit note or re-issued invoice from the source, not a patched PDF. For your own draft invoices, replace freely before sending - after sending, the document is executed.

Quoted proposals and contracts in revision. Replacements between versions are normal - and the version discipline is what keeps them legitimate: new filename per revision, a change log line in the email, and never editing a countersigned version (the addendum workflow governs there).

Regulatory and compliance filings. Submitted documents are records; corrections flow through the filer's amendment process. A patched PDF filed alongside an unpatched original in a regulator's system is an inconsistency finding regardless of intent.

Academic and journalistic texts. Post-publication corrections run through corrigenda and editor's notes by convention - silently replacing a published figure breaks the versioning that scholarship and journalism depend on.

Personal records. Statements, certificates, medical records: these belong to their issuing institutions. Any replacement - however innocuous the intent - creates a doctored document, which is the forgery line from the signed-PDF guide. Requests for corrections go to the issuer.

The common thread: the more a document is relied upon, the more its changes must happen in systems of record rather than in pixels. Cover-and-retype is for working documents; institutions amend their own records.

A quality-control pass for replacements

Professional document shops run a two-minute QC pass after batch replacements - worth adopting whenever more than a handful of edits accumulate:

1. Inventory what you changed. Before starting, list every replacement (page, old, new). After finishing, walk the list and confirm each is done - the list catches the "I'll remember" instances that slip.
2. Search for leftovers. In the saved file, search for the old term. Method 1 will find the hidden originals (expected); visible leftovers - instances you missed - are the actual catch. Cross off found-and-fixed items mentally as you locate them.
3. Zoom sweep. One pass through affected pages at 200%: every cover and replacement gets two seconds of scrutiny. This is where tilted baselines and peeking glyphs reveal themselves.
4. Print test for critical pages. Any replacement on a page destined for print gets one test print - screen-perfect is not print-perfect, and the affected pages are known, so the test is one sheet, not one ream.
5. Version stamp. Update the document's Title metadata (the metadata editor takes seconds) - "Proposal v3" instead of v2 - so inbox copies are self-identifying.

The QC pass converts replacement work from "I think I got them all" to "here is the list of what changed" - which, in any context where someone might ask, is the difference between confidence and hope.

Replacement across document families

Real replacement jobs often span more than one file - a renamed product across a spec sheet, a rebranded term across a template library - and the family context changes the method.

Across a handful of related files: the Word route per file, with the same find-and-replace list applied to each, is the honest approach - and the QC pass's inventory list becomes a simple matrix (files down the side, terms across the top) so no file-term combination gets missed. Two hours of focused work rebrands a twenty-document set this way.

Across a template library: fix the templates, not the exports. If the PDFs regenerate from templates (Word, design tools), the replacement happens once in the template and every current and future document inherits it. Editing exported PDFs in a template library is painting over the well - the wrong term returns with every regeneration.

Across versions of the same document: resist the temptation to patch the newest version and call it done; the version chain needs the change noted (version stamp, change line) or the next reader cannot tell patched-v3 from true-v3. The metadata editor's Title field plus a one-line email is sufficient provenance for most teams.

When many hands are involved: the family job needs one owner of the replacement list - distributed "everyone update their docs" produces exactly the drift you would expect. A single operator with the matrix finishes faster than a committee, and finishes consistent.

The family scale also decides the fidelity question: with dozens of files, the Word route's rebuild becomes desirable (clean text layers, no hidden originals), where on a single precious document it might be prohibitive. Scale flips the trade-offs.

How desktop editors replace text (and why results vary)

For completeness, the paid-desktop route deserves its honest paragraph - including why its results range from seamless to mangled.

Acrobat's Edit PDF mode and equivalents (Foxit, Nitro) perform true structural replacement: click into a text block, retype, and the editor rebuilds the internal text runs - substituting your font if the original is unavailable, and reflowing the block as needed. With the original font installed and short replacements, results are excellent. The three failure modes, all font-driven: missing fonts trigger substitution that changes letter widths (neighboring lines then re-break); heavy edits inside justified paragraphs produce visibly uneven word spacing; and text converted to outlines or produced by exotic tools imports as shapes that no editor can retype.

Two production notes if you hold an Acrobat license: check File > Properties > Fonts before editing and install what is missing - this single habit prevents most bad outcomes - and prefer many small edits over one large rewrite, because block-reflow errors scale with the edit size.

The free-vs-paid calculus from this guide's perspective stays honest: desktop editors are the right tool when you already pay for them and the original fonts are at hand. For everyone else, cover-and-retype achieves the same visual result for isolated changes, and the Word route out-replaces any desktop editor for document-wide ones - at zero cost, with the same local privacy.

Copywriting-adjacent replacements: names, terms, and numbers

Certain replacement categories carry their own conventions - worth thirty seconds each because they come up constantly.

Person names. Replacing a name (marriage, correction, transliteration change) also requires catching titles and possessives: "Ms. Alvarez" becomes "Ms. Chen", and "Alvarez's" needs its possessive handled - find-and-replace with whole-word matching misses possessives, so run a second pass for the name plus apostrophe. In correspondence, a one-line note of the correction preserves courtesy.

Product and feature names. Marketing renames have case conventions (iPhone, PayPal, the internal "Project Falcon" to "Skyline") - find-and-replace is case-sensitive by default in most tools for good reason, and a case-insensitive pass afterward catches shouting remnants in headings.

Dates. The year swap ("2025" to "2026") is the classic January edit - and the classic partial-edit trap, because date ranges ("2024-2025"), copyright lines, and file-path references each need individual judgment. Replace, then scan every hit's context rather than trusting blanket replacement.

Prices and figures. Numbers carry meaning in totals: replacing a line-item price without the recalculated total creates internal inconsistency that any reader with a calculator finds. Fix the arithmetic together or not at all - and in invoices, prefer re-issue (per the scrutiny section).

Legal references and clause numbers. Renumbering a contract's clauses shifts every cross-reference - this is edit-architecture work that belongs in the source document with automatic cross-references, not in pixel patches. If a clause number must change on a PDF, search the document for every reference to it before declaring the job done.

Each category's lesson is the same: replacements ripple, and the ripples live in context - the find is easy, the sweep for consequences is the work.

The hidden-text inventory: what else your PDF is hiding

Cover-and-retype leaves originals hidden under covers - and covers are not the only hiding places. A document's true content often exceeds its visible content, which matters both for recipients' privacy and your own due diligence.

The hiding places, inventoried:

• Covered text (your Method 1 originals) - searchable, extractable, present.
• Annotations and comments - review threads, sticky notes, and markup history that "look clean" in full-page view.
• Form field values - typed entries persist in the file structure even after visual clearing.
• Deleted-but-recoverable content - incremental saves mean "deleted" pages and text sometimes survive as unreferenced objects in the file.
• Document metadata - author names, revision history, application fingerprints, and sometimes path names revealing folder structures.
• Embedded attachments and scripts - files-inside-files and, rarely, JavaScript actions from untrusted sources.

The due-diligence pass before sharing anything sensitive: run the Sanitize PDF tool - it strips metadata, annotations, embedded data and hidden text in one local pass - then search the sanitized file for every sensitive string you know about. Zero results across the board is the ship condition. The same pass is worth running on received documents you forward: the previous sender's comments travel with the file unless you remove them.

This inventory reframes Method 1's "limit" as a manageable property: covered text is fine for layout corrections on non-sensitive content, and the sanitize pass closes the gap everywhere it matters. Knowing which of your documents contains which - that is the operational skill.

FAQ

How do I replace text in a PDF for free? Cover the old text with a background-colored rectangle and type the replacement in the free browser-based Edit PDF tool - or convert to Word for free and use find-and-replace when replacements are numerous.

Why is there no simple find-and-replace in PDF editors? PDF text is positioned glyphs split across internal runs, not flowing sentences - there is no coherent string to search. Word-based replacement sidesteps the architecture entirely.

Does covering text with a white box really remove it? No - it hides it. The original remains in the file for search and extraction. True removal requires redaction, which deletes the text object itself.

How do I replace text in a scanned PDF? Cover the old text with a scan-matched shape and type over it. For many replacements, OCR first, then convert to Word and replace properly.

Why does my replacement text look different from the surrounding text? Check the four matches in order: font category, visual size (x-height), color (sample the original - usually near-black), and baseline position.

Can I replace the same word everywhere in a PDF at once? Yes, via the Word route: convert to DOCX for free, Ctrl+H replaces every instance, then export back to PDF.

Will the replaced text be selectable and searchable? Your new text is real selectable text - and so is the covered original, which is why sensitive replacements must be redacted first.

Is it legal to replace text in a PDF I did not create? Authorized edits - drafts, your copies, agreed revisions - are normal. Changing the meaning of contracts, statements or records you do not own is forgery.

Replace your text now

Open the free Edit PDF tool for cover-and-retype, or the PDF to Word / Word to PDF pair for find-and-replace - all local, all free. Sensitive originals? Sanitize PDF strips what hiding leaves behind.