Plenty of PDFs were never built as digital forms. A scanned application from a doctor's office, an old government form, a landlord's custom lease: these have no AcroForm fields, so clicking them does nothing. When you upload a flat PDF, the AI analyzes the page image and proposes positions for text inputs, checkboxes, and signature areas based on labels, underlines, and box outlines.
Once detection finishes, the proposed fields appear as overlays on top of the page. You click into a spot and type, just like a regular form. If the AI missed a field or placed one slightly off, you can drag, resize, or add new ones manually. Manual overlays work on any PDF, so you are never blocked by detection gaps.
Detection accuracy depends on the source document. Clean scans with clearly labeled fields tend to work well. Dense, heavily designed, or low-contrast pages may need more adjustment. The AI feature requires an account (20 free starter credits on signup); manual text overlays are free and do not require signup.
Not currently. When you download, the typed content is baked into the PDF as flattened text on the page rather than re-saved as AcroForm fields. If the original PDF already had AcroForm fields, those values are preserved in the form structure on save.
It varies with the source. Standard forms with printed labels and underlines or boxes tend to detect cleanly. Scans with skew, noise, or unusual layouts may need you to nudge positions or add fields by hand after the pass completes.
The AI can detect checkbox positions and place a toggleable overlay at each one. Radio groups (mutually exclusive options) are not grouped automatically yet, so each option behaves as an independent checkbox you toggle yourself.