Edit PDF Text on Android

Step by step

These steps are the current build running on a phone. Every screenshot below is a real capture from the same handset, over the same sample contract, with the app in the language of this page.

1

Open the PDF

Pick the file in the system picker. While it opens, PDFMason diagnoses the document: page types (native text, scanned, mixed, vectorized, empty), hidden OCR layers, and a five-tier editability grade with a specific reason. You know what you are dealing with before the first edit.

2

Select the text

Tap a block to select it whole. Long-press a word for a character-level selection, then drag the handles to extend it — across style changes if needed. The selection can be copied to the clipboard.

3

Type the replacement

The editor removes the original glyphs, shapes the new run, and writes it back with a re-embedded font subset. Font family, size and color are adjustable for the edited run. The paragraph stays one real text object, not two overlapping ones.

4

Resolve overflow

If the replacement does not fit the frame, choose one of three outcomes: shrink the font to a suggested scale that fits exactly, truncate the text, or expand the frame.

5

Handle a refusal, if one comes

If the embedded font cannot draw a character you typed, the edit is refused before you invest work, and the offending characters are named and flagged. Switch to a replacement font from the picker and the same edit writes cleanly.

6

Save and check the report

Save as a copy or in place. The app verifies the file after writing and shows what changed. Undo history and a crash-safe draft back the whole session.

What you do not get: a new text box stacked over the old line, two searchable copies of the same sentence, or a rasterized picture of your paragraph. The underlying mechanics — glyph arrays, font subsetting, why most editors skip this — are explained in Edit existing text in a PDF.

When it works and when it doesn't

PDFMason grades every document into one of five editability levels when it opens. The grade is shown, not guessed, and each level means something different for your edit:

Two honest notes. Overlay-only documents are labeled rather than faked: v1 has no overlay editing mode that paints new text over the outlines, so the app shows the label instead of pretending. And even a directly editable document can refuse individual characters — that is a font coverage problem, covered in the next section.

Known limitations, stated plainly

The product-wide limitation list, in the same words used everywhere on this site:

Password-protected files, browser versus app: “The browser tools can't open password-protected PDFs. The Android app can.”

The browser side is deliberately smaller than the app: “No OCR, no compression, no PDF-to-Word in the browser — yet. We'd rather ship nothing than ship a version that doesn't work.”

Checkbox and radio form fields: “Checkbox and radio-button form fields are marked unsupported in the app. We can change them in memory but the change doesn't survive saving, so we don't pretend otherwise.”

Non-Latin form filling: “Filling non-Latin text fields can produce a file that appears blank in other readers. The app warns you per field.”

Line and arrow annotation shapes were tested on real documents and cut: “We tested it on real documents: it only writes correctly to 33% of them. So it's not in v1.”

Saving very large documents: “A 1,136-page document takes 24–32 seconds to save with verification. It runs in the background and you can cancel it.”

Specific to text editing: there is no document-wide find and replace, no Indic scripts, no vertical text, and no global reflow after an edit. An edit changes the paragraph you touched, not the pagination of the whole file.

Fonts and why text sometimes can't be written

A PDF does not store the letter A. It stores a glyph ID drawn from an embedded font program. To keep files small, producers usually embed a subset: only the glyphs the document actually uses. If a page says “Quarterly review”, the embedded font may contain Q, u, a, r, t, e, l, y and the space — and nothing else. Type “Zebra” into that line and there is no Z in the font to draw it with.

That is why text sometimes cannot be written back, and it is the failure mode most editors hide: they draw a black box, substitute a wrong glyph, or silently drop the character. PDFMason does the opposite. Font coverage is checked before the edit is applied; when it fails, the edit is refused and the app names the exact characters it cannot write, flags them with a red wavy underline, and disables saving until the conflict is resolved. A font picker offers replacements, and the font you switch to is subset and embedded on write-back, so the file stays self-contained.

Shaping — turning characters into positioned glyphs — is handled for Latin, CJK, Arabic and Hebrew scripts. Two honest edges: Indic scripts and vertical text are not supported in v1, and there is a known defect in fallback font selection where a purely Latin run can be won by a CJK fallback font. Scanned pages have no text at all until OCR runs; after that, OCR quality, not the editor, sets the ceiling on how correct the recognized text is.

FAQ