PDF Guides & How-Tos

Practical, plain-English guides for everyday document work — scanning, reading text from scans, shrinking files for email, and reorganising pages. Every method below runs entirely in your browser: your files are never uploaded to a server.

How to scan a document with your phone

A phone camera can produce a scan that is as readable as a flatbed scanner — the difference is almost entirely in lighting and framing, not megapixels.

Before you shoot

• Use indirect light. A window to the side beats a lamp overhead, which throws your own shadow across the page.
• Put the page on a contrasting surface. A white page on a dark desk gives clean, findable edges. A white page on a stack of other white papers gives none — automatic edge detection will struggle, and so will you.
• Fill the frame, but keep all four corners visible. Cropping tightly in-camera wastes resolution on nothing; letting an edge run off the frame makes straightening impossible.
• Hold the phone parallel to the page. Shooting at an angle turns a rectangle into a trapezoid.

Then

1. Open Pennary on your phone and tap the camera button.
2. Capture the page. The crop box appears over it automatically.
3. Drag the corner dots and side bars so the box hugs the paper — this is how you cut out fingers, desk, and shadows.
4. Cycle the enhance button through Off → Auto → Grey → B&W and pick what reads best (see below).
5. Tap the tick to add the page, then add more pages or finish.

Which enhance mode?

• Off — the raw photo. Best when colour matters and lighting was already even.
• Auto — evens out lighting while keeping colour. Best default for documents with stamps, seals, or signatures in blue or red ink.
• Grey — same correction without colour. Smaller files, very readable.
• B&W — pure black and white. Crispest for plain text you will photocopy, but harshest: faint pencil and light print can drop out entirely.

How to extract text from a scanned PDF (OCR)

There are two completely different kinds of PDF, and knowing which you have saves a lot of time.

Born-digital vs scanned

A born-digital PDF was exported from Word, a browser, or an accounting system. The text is real text — already selectable and searchable. A scanned PDF is a photograph of paper: to a computer it is just pixels, with no text in it at all. Optical Character Recognition (OCR) is the process of reading those pixels and working out the letters.

Quick test: try to select a line of text with your cursor. If it highlights, it is born-digital and you need no OCR — just copy it.

Steps

1. Open the PDF in Pennary.
2. Choose Read text (OCR).
3. If the page already has real text, it is extracted exactly — instantly and perfectly.
4. If it is a scan, the page is read with OCR and the recognised text appears in an editable box.
5. Correct any mistakes in the box, then copy the text.

Getting better OCR results

• Resolution matters more than anything. Small, blurry text is guesswork for any engine. Fill the frame with the page.
• Straighten first. OCR reads along horizontal lines; a tilted page lowers accuracy sharply.
• Try Grey or B&W. Clean, high-contrast strokes recognise better than a grey, uneven photo.
• Expect to proofread. No OCR is perfect, especially on stamps, handwriting, tables, and unusual fonts. Treat the output as a strong draft, not a finished document.

How to compress a PDF so it fits an email

Most mail servers reject attachments over roughly 10–25 MB. Scanned PDFs blow past that quickly, because each page is a full photograph.

Why scans are so large

A born-digital PDF stores text as instructions — a few kilobytes a page. A scan stores millions of pixels. Ten scanned pages at high resolution can easily reach 40 MB, while the same document exported from Word might be 200 KB.

Steps

1. Open the PDF in Pennary and choose Compress PDF.
2. Pick a quality level. Lower quality means a smaller file and softer text.
3. Check the result before sending — zoom in on the smallest text on the page.

Practical advice

• Compression is lossy. Detail removed to save space does not come back. Keep your original.
• Scanning in Grey or B&W shrinks files at the source, often more effectively than compressing a colour scan afterwards.
• Split instead of squeezing. If a 60-page file will not fit, sending pages 1–30 and 31–60 as two readable files beats one unreadable one.
• Do not over-compress anything official. A tender document or certificate that arrives illegible costs more than a second email.

How to convert PDF pages to images

Useful when you need a page inside a slide deck, a report, a chat message, or anywhere a PDF will not embed.

Steps

1. Open the PDF and choose Convert / Export.
2. Select an image format.
3. Download the pages.

JPG or PNG?

• JPG — much smaller, ideal for photographs and scans. Its compression slightly blurs sharp edges, which can smudge fine text.
• PNG — larger, lossless, keeps text edges crisp. Better for pages of text, diagrams, tables, and line drawings.

Rule of thumb: a photograph of a page → JPG. A page that is mostly text or a diagram → PNG.

How to merge, split, and rotate pages

Merging several PDFs into one

1. Choose Merge PDFs.
2. Select the files. Order matters — they are combined in the order given.
3. Download the combined file.

Common use: an application form, its annexures, and supporting certificates arriving as one clean document instead of six attachments.

Pulling out the pages you need

1. Choose Split / Extract.
2. Enter the pages — for example 1-3, 5, 8.
3. Download just those pages as a new PDF.

Useful for sending one relevant page of a long report, or removing pages you should not circulate.

Fixing orientation

Pages scanned sideways or upside down can be rotated individually or all at once. Rotate before running OCR — text recognition depends on the page being the right way up.

A note on privacy

Every guide above describes work done on your own device. Pennary processes PDFs in your browser using open web technology; your documents are not uploaded, stored, or seen by us — which is the point. For sensitive material — contracts, medical records, government files, anything with personal data — that difference matters more than any feature.