Snap a Paper Form, Get a Mobile Form: AI Digitizes the Fields for You

The Problem: Rebuilding Paper Forms Field by Field

Most field crews already have a paper form that works. The inspection checklist printed on the clipboard. The site survey from a folder. The QC pass/fail sheet someone designed in Word three years ago. It's been refined, the fields are in the right order, and everyone knows it.

So when you switch to a mobile form builder, the friction isn't the *concept* of going digital — it's the half-hour of clicking *Add field → Pick type → Type the label* twenty-five times to recreate something you already have on paper.

That's where most trial users give up. The form they want is sitting on the desk in front of them. They just don't have the patience to retype it.

Paper form import removes that step. Snap a photo. Get a mobile form.

How It Works

Three steps. Most of the work happens in the first one.

1. Tap "Start from a paper form." On the *Create a new form* dialog, pick the featured option. (You can also upload a single-page PDF if you have a scan saved.)

1. Pick a file. Take a fresh photo with your phone, or upload an existing image (JPG, PNG, HEIC) or single-page PDF up to 10 MB. We compress and resize on your device before uploading — your bandwidth and battery don't take a hit.

1. Review and save. A few seconds later you're in the form builder with the title pre-filled, every field typed, and a banner reminding you to double-check. Fields the AI wasn't fully sure about have a small yellow dot — your cue to look closer.

That's it. No mapping, no template hunting, no retyping.

What Gets Recognized

The AI maps each visible field on the paper form to the closest MobileForm field type. The mapping rules, in priority order:

• Yes/No, Pass/Fail, OK/Not OK, ✓/✗ → *Pass/Fail* field
• Single checkbox or "I agree" → *Checkbox* field
• Multiple-choice with one selection ("circle one", radio buttons) → *Dropdown* with the printed options
• Numeric measurements, counts, scores → *Number* field
• "Date:", "Inspection date" → *Date* field
• "Time:", "Inspection time" → *Time* field
• "Signature:" line → *Signature* field
• "Photo:" or "Attach picture" → *Photo* field
• GPS or coordinates → *GPS* field
• 1–5 stars or scale ratings → *Rating* field
• Multi-line write-in space → *Long text* field
• Everything else → *Short text*

Headings, instructions, branding, and pre-printed signatures are deliberately ignored — those aren't fillable fields, and digitizing them as input boxes would just clutter the form.

Where the AI Gets Less Sure (and How You'll Know)

Vision models are very good at typed paper forms with clear field labels. They're less good at:

• Heavily handwritten or annotated forms
• Multi-column layouts where it's not obvious which question goes with which answer space
• Multi-select questions ("check all that apply") — these collapse to a single-select for now
• Heavily branded or stylized layouts where labels and decoration blur together

For any field where the AI guessed the type from context rather than an explicit cue, the field card in the builder shows a small yellow dot. That's not an error — it's a "please double-check this one" signal. Most of the time the type is right; sometimes you'll want to swap a *Short text* for a *Number*, or fix a label.

Save the form when it looks right. The yellow dots don't get persisted; they're a one-time review hint.

Limits

A few things to know up front:

• Single-page only. Multi-page PDFs are rejected with a clear message — export just the form page and try again. We chose this on purpose: most paper field forms are one page, and trying to merge a multi-page PDF into one digital form usually produces confusion, not value.
• No image storage. The photo or PDF is processed and discarded. We don't save the original — only the extracted fields and your eventual edits. If you want to import the same form twice, you'll need to upload it again.
• Monthly quota by plan. Free includes 3 imports per month. Pro includes 50. Team is unlimited. Failed extractions (the image wasn't a form, or no fields could be read) don't count against your quota.

If you have a paper form right now, give it a try. It's the fastest way from "I should digitize this" to "this is done."