How to Scan Documents with Your Phone (and Get a Usable Result)

Phone cameras have been good enough to replace a flatbed scanner for years. The difference between a scan that looks professional and one that looks like a photo of some paper comes down to four things, none of which require any special app.

1. Light, which matters more than the camera

Poor lighting causes more unusable scans than anything else. Three rules:

• Diffuse, even light. Near a window on an overcast day is close to ideal. Direct sunlight creates harsh shadows and blown-out highlights.
• Avoid your own shadow. The single most common mistake — standing between a ceiling light and the document. Light from the side, or move the paper.
• Do not use the flash. It produces a bright hotspot in the centre and dark corners, and on glossy paper it reflects straight back.

2. Geometry: flat, square and parallel

Hold the phone parallel to the page, not at an angle. Photographing from an angle produces keystone distortion, where the page becomes a trapezium. Some apps correct it; the correction always costs sharpness.

Flatten the document. Fold creases and curling cause both distortion and uneven lighting. A book held open at the spine will have a dark, curved gutter — press it flat or photograph pages separately.

Fill the frame, leaving a small margin. Cropping in afterwards throws away resolution you may need.

3. Focus and stillness

Tap to focus on the text before capturing — phone cameras frequently focus on the surface behind the document. Then hold still; indoor shots use slower shutter speeds and motion blur is the second most common cause of failed OCR.

Bracing your elbows on a table, or resting the phone against something, costs nothing and helps considerably.

4. Background and contrast

Place white paper on a dark, matte surface. The contrast lets edge detection find the page boundary reliably, and makes automatic cropping far more accurate. White paper on a white desk is the worst case.

Turning the photo into a document

Once you have good captures, the workflow is short:

• iPhone users — convert the HEIC files first with HEIC to JPG, since many tools will not read HEIC.
• Combine the pages into one document with Image to PDF, in the right order.
• Fix orientation if any page came out sideways, using Organize PDF.
• Make it searchable with OCR PDF, which adds an invisible text layer so the document can be searched and copied.
• Compress with Compress PDF if it needs to be emailed.

Getting the best OCR results

If the scan will go through optical character recognition, these matter disproportionately:

• Resolution. Fill the frame with the page. Roughly 300 DPI equivalent is the practical floor; below about 200 accuracy drops sharply.
• Straightness. OCR engines correct a few degrees of skew, not perspective distortion.
• Contrast. Crisp black on white. Faded receipts and grey photocopies are genuinely hard.
• Language. Select the right one. Running English recognition over a French document produces nonsense.

A note on sensitive documents

People photograph passports, bank letters and medical paperwork constantly. Those images usually then get uploaded to whatever free converter appears first in search results.

Every step in the workflow above runs entirely in your browser — the photographs and the resulting PDF never leave your device. For documents of that kind it is worth the small effort of choosing tools that work this way.