Use Everyday Tools

HOW-TO GUIDE · BROWSER-FIRST WORKFLOWS

Edit existing PDF text without turning the page into an image

Understand true text-run replacement, embedded-font limitations and the difference between editing and annotation.

Editing and annotation solve different problems

Adding a note or new text on top of a page is annotation. Replacing existing text means changing the underlying text content. A white rectangle and new lettering can look correct while leaving the original words extractable. A rasterized export removes the page text layer entirely. Choose a workflow based on what the recipient needs to search, copy or edit later, not only on how the preview looks.

Three kinds of “PDF editing” that behave differently

The phrase covers very different techniques, and they leave different traces. The first draws a white box over the old words and types new words on top, so the old text is still inside the file and can be found by search or copy and paste. The second converts the page to a picture and edits the picture, which removes the text layer altogether. The third, which this tool uses, locates the text in the page’s content and replaces the characters, so the page stays vector, selectable and searchable.

Which you want depends on what happens next. If the recipient will search or copy the text, avoid the picture approach. If you are trying to hide something, none of the editing approaches is a reliable way to do it; that needs redaction, covered at the end of this guide.

What this editor changes

For supported page text runs, the editor rewrites text-showing operands and reuses the original font. Other content on the page is not intentionally rasterized. The run list is deliberately conservative: unsupported encodings, text inside Form XObjects and some inline-image streams are not offered. This is not a complete desktop editing engine and it does not reconstruct a word-processing document.

The kinds of change that work well

The editor is at its best for small, factual corrections: a misspelled name, a wrong date, a changed invoice or reference number, a corrected amount. These replace a short run of existing text with a similar amount of new text, using the characters the document’s font already contains.

It is a poor fit for rewriting a paragraph, adding a new line of text or changing the layout. Nothing reflows, so a longer replacement can overlap its neighbors, and a shorter one leaves a gap. For heavier changes it is usually quicker to convert to Word, edit there and export a fresh PDF, accepting the layout clean-up that involves.

Why some replacement characters fail

Embedded subset fonts often contain only the glyphs originally used. A replacement digit can work while a new accented letter fails because the glyph is not present. The editor rejects that substitution rather than drawing a different font without warning. Keep edits short, use available characters and choose another application when you need a larger wording change or new language.

Fonts: why a letter sometimes cannot be typed

Most PDFs embed only the part of a font they actually use, called a subset, to keep the file small. If the original text never contained an “é” or a “7”, the subset has no shape for it, and the editor cannot draw one without substituting a different font. The tool refuses rather than quietly mixing fonts, which would look wrong and might not match when the file is reprinted.

In practice this means numbers and letters that already appear in the document are usually available, and unusual characters often are not. If you need a character the font does not contain, regenerate the PDF from the source document, where the full font is available.

A safe correction workflow

Use an unsigned original. Select the page and run, apply the replacement and preview the actual prepared PDF. Check spacing around neighboring words and numbers; the editor does not automatically reflow paragraphs. Download a separate copy, search for the replacement and copy the line into a plain-text editor. Confirm that the old wording no longer appears in the selected run and that surrounding content is intact.

Checking the result properly

A good-looking page is not proof that the change was made correctly. Do these checks every time, using a second viewer that is not the one you edited in.

  • Zoom to the edited line and check spacing against the words on either side.
  • Search for the new wording; it should be found. Search for the old wording; it should not appear in that line.
  • Select the line and paste it into a plain-text editor to confirm the characters are real text, not shapes.
  • Check page count and that other pages did not change.
  • Save the edited file under a new name and keep the original unmodified.

Do not use ordinary editing as redaction

Removing a visible text run is not a guarantee that all copies of a secret are gone. Sensitive values may also exist in metadata, attachments, form fields, layers or duplicated objects. Use a dedicated redaction process and verify text extraction afterward. For legal or compliance documents, follow the required review process rather than relying on a visual correction alone.

When you cannot edit and what to do instead

The editor declines scanned pages, signed PDFs and text stored in encodings it cannot interpret safely. A scanned page is a picture, so there is no text to edit; run OCR first and recreate the document if the content must change. A signed PDF cannot be modified without breaking the signature; ask the signer for an unsigned source or issue a corrected, newly signed version.

If the tool offers no text run where you expect one, that is information: the text may be in an unsupported structure such as a form field or an image. In that case, check whether the document is a form (use the form tools) or ask for the original file.

Questions people ask

Why can I edit some lines but not others? The editor offers text runs it can change safely. Text in form fields, images, unsupported encodings or special structures is not offered.

Does editing change the font? The replacement uses the original embedded font. If that font lacks a character, the tool declines instead of substituting a different font.

Can I edit a scanned PDF? No. A scan is a picture. Run OCR first and recreate the text, or ask for the source file.

Will my edit keep the file small? Yes. The page is not converted to an image, so size stays close to the original.

Keeping a record of corrections

When you correct a document that others rely on, such as an invoice, a certificate or a report, record what you changed. A one-line note in your own files is enough: the date, the page, the old and new wording and who requested it. If the corrected file ever has to be explained, this note prevents a dispute about what was altered.

For documents with legal, financial or regulatory weight, a quiet correction is not always appropriate. Issuing a new version with a clear revision note, or asking the original author to reissue it, is often the right course. A tool can make the change; it cannot decide whether the change is yours to make.

Before you finish

  • Keep an untouched original and use a separate filename for the result.
  • Test a representative sample before processing the full document.
  • Review the actual downloaded file in the receiving app, not just a preview.

← More practical guides · Report an error or suggest an improvement

Try a known sample first

Use synthetic files with expected page counts, text and form values. Compare the actual downloaded result, not just the preview.

Worked examples and sample files · Testing approach and limitations