PDF filling: fields or placed text?

Start with the right file and operation

Smallpdf’s Fill Form screen identifies files by name, date and size and shows one selected file before Continue. PDF Fill & Sign keeps the file name visible above the document and exposes annotation tools below it. These states answer two different questions: which document is about to be opened, and what operation is currently available on that document.

Retain source identity throughout the task. Similar file names, a compressed copy and a revised attachment can all look alike in a recent-files list. Show a page preview or useful metadata before editing, and keep the chosen file’s identity available once the keyboard covers half the page. Do not automatically overwrite another version merely because the names match.

When a PDF is locked, unsupported or read-only, explain the limitation before accepting a page of input. Offer a safe copy workflow only when it is supported. If a document contains real fields, use their labels and structure where possible; if it does not, present text-placement mode honestly. A broad Fill Form label can be an entry point, but it should not promise automatic field detection for every file.

Distinguish adding text from editing source text

pdfFiller displays Text and Edit text as separate tools. PDF Editor: Read Scan Convert instructs the person to select a boundary to edit or tap where text should be inserted. The distinction deserves deliberate wording: a new label floating over a page is not the same as changing an existing paragraph in the source document.

Use a persistent mode cue, a clear selected object and a direct exit. For placed text, a tap defines a new object and its location; for a real field, it changes a value within that field. Do not let the same tap unpredictably switch between those meanings based on invisible recognition. If automatic detection is offered, show the detected target and let the person override it before typing.

Mode errors can leave documents looking correct at one zoom while hiding duplicated text underneath. Test by selecting, reopening and exporting the result, not just by viewing a screenshot. Keep the undo operation local to the intended object, and make deletion remove that object rather than the document page. The public interface should never imply that visual covering removes underlying source content.

Preserve location when the keyboard appears

pdfFiller’s active text box has handles and formatting controls while the keyboard is open. Forma shows a separate text input strip with Add, Autofill and a date shortcut. These represent different input models: editing directly on the page and entering content in a dedicated composer before placement. Either can work if the resulting object and commit action remain identifiable.

Keep the active field or insertion point above the keyboard and distinguish its border from the PDF’s printed lines. Preserve the viewport when the keyboard closes; a sudden jump to the page top makes it difficult to confirm alignment. Avoid shrinking every page to a tiny full-sheet view just to keep a decorative toolbar visible. The text and target need priority over unrelated tools.

Autofill should expose which stored value it inserts and allow editing before commitment. Do not silently reuse personal details from another document or send a document to a service without the appropriate disclosure and authorization. Treat copied text, long names, multiline addresses and right-to-left input as normal cases. Keep composition text intact when switching to formatting or closing a temporary tool.

Make coordinates and appearance reviewable

MobiPDF prompts Tap where you want to add text within Fill & Sign. Its Text Box panel separately exposes Text and Shape, fill color, opacity and border. That separation matters when a document has narrow printed boxes: a background fill or border can obscure information even though the characters themselves are correct.

Store placement in document-page coordinates, not viewport pixels. Zooming, rotating the device and opening the keyboard should not change the exported position. Give a selected object enough contrast and handles to adjust without accidentally creating another object. Provide accessible alternatives for moving or resizing when precise dragging is difficult, and preserve the text while those controls are used.

Keep visual styling subordinate to legibility and the document’s purpose. A decorative font picker is less important than fitting the complete value into the intended area. Warn about overflow or clipping rather than silently shrinking text until it cannot be read. Where structured fields have character limits or required formats, explain those constraints next to the actual field and preserve the rejected input for correction.

Review marks and the actual saved artifact

PDF Fill & Sign exposes Change to Crossmark and Remove for a selected checkmark, with an instruction to tap anywhere to add one. PDF Editor shows Completed inside an active text-entry surface. That word is entered document content, not evidence that the file has finished saving. Interface status and document text must never be conflated.

Before export, provide a clean review surface without handles covering the values. Check every page, especially required fields below the fold. A completion action should identify the saved copy and its destination, while a background save should report failures without discarding the entered values. Do not claim a document is submitted or accepted by its recipient when the app has only created a local file.

The resulting artifact then enters a sharing and export flow. Test reopening that artifact in another supported viewer and verifying text, placement and checkbox behavior. State whether values remain editable or are flattened only when the implementation confirms it. No screen in this comparison establishes legal validity or recipient acceptance.

Test one document in more than one viewer

Build a fixture set with a structured form, a scanned form image and a plain PDF page. Complete the same types of values in each: a short name, multiline address, non-Latin text and a checkmark. Save, close and reopen the actual generated artifact in another supported viewer. Verify position, legibility and editability separately; a successful preview is not proof of all three.

Include a page rotation and a zoom change during entry. If text jumps after export, inspect the mapping between screen coordinates and PDF coordinates rather than compensating with a visual offset that only works on one device. Keep the source revision and page identity attached to each placed object, especially when pages can be inserted or reordered while a document is open.

Avoid putting document contents in analytics. Record failure categories such as placement-save failure or unsupported field type without copying personal values. In usability review, ask users to locate the final copy and explain whether it is still editable. Their understanding of the artifact is part of completion, not an optional detail after the editor closes.

Test beyond the editor preview

Use these checks on a working build, not only on the polished reference state.

• ✓Identify whether the PDF has structured fields, page images or ordinary source text.
• ✓Keep add-text, edit-source-text and field-value operations distinguishable.
• ✓Preserve active content and page location across keyboard, zoom and orientation changes.
• ✓Store overlays in page coordinates and test multiline and right-to-left values.
• ✓Make checkmarks, crossmarks and their removal unambiguous without implying data-field support.
• ✓Verify every page in the saved output and reopen it in another supported viewer.
• ✓Explain the destination, editability and failure state without claiming successful submission.