You scan a document to PDF by capturing it with a scanner, phone camera, or multifunction printer, then saving the output as PDF rather than JPEG or TIFF. The single setting that decides whether the file is usable is not resolution — it's text recognition, which most default scan profiles leave switched off. Turn on OCR (or "Searchable PDF" mode) before you press scan and you get a document you can search, copy from, and file; skip it and you get a picture of a page that no one can find by keyword later.
What the PDF Format Actually Stores When You Scan
A scanned PDF is a container holding one image per page, plus optional layers. The image layer is the scan itself. The text layer, added by OCR software, sits invisibly on top and makes the page searchable. A third layer holds metadata — creation date, device, page count — which is why two files of identical size can behave completely differently in a document management system.
The format is defined by ISO 32000, maintained by the International Organization for Standardization, so the container itself is stable across every scanner and app you'll meet. What varies is how each device populates the layers. A $150 flatbed and a phone app both produce valid PDFs; only one of them is likely to contain a text layer by default.
PDF/A is the archival subset of the format, and it's the version to pick when a document must remain readable in twenty years. It forbids embedded audio, video, and encryption, and requires all fonts to be embedded. Government recordkeeping rules in the United States, under 36 CFR Part 1236 from the National Archives, and in the European Union, under national implementations of archival standards, commonly point to PDF/A for permanent retention.
The practical distinction for a home or small-office user comes down to three formats. Standard PDF is fine for email and filing. PDF/A is what you choose for deeds, certificates, tax records, and anything a lawyer or archivist might ask about. Searchable PDF is a standard PDF with an OCR text layer and is the sensible default for contracts, invoices, and receipts.
Step 1: Match the Scanner to the Document Type
Choose the capture device before you touch settings, because the device determines the maximum resolution and whether automatic document feeding is available. A flatbed scanner handles bound books, passports, and creased receipts but requires you to place each page. An automatic document feeder, standard on office multifunction printers, takes 25 to 50 sheets at a time and scans both sides in one pass. A phone camera wins on portability and loses on geometry.
| Document type | Best capture device | Typical resolution | Why |
|---|---|---|---|
| Single contract page | Flatbed scanner | 300 dpi | Flat, no motion blur, OCR-friendly |
| 30-page mortgage packet | ADF multifunction printer | 300 dpi | Duplex pass, consistent alignment |
| Receipt or business card | Phone camera app | Native camera output | Always available, no setup |
| Bound book chapter | Overhead or flatbed with edge | 300–400 dpi | Avoids spine shadow |
| Handwritten form | Flatbed scanner | 300–400 dpi | Highest OCR accuracy for cursive |
| Photo print to share | Phone camera | Native camera output | Color fidelity matters more than text |
Scanning at 300 dpi is the standard recommendation for text documents, and 600 dpi is the usual figure for documents containing small print or handwriting you intend to run through OCR. Going above 600 dpi on plain text adds file size without adding readable information. Going below 200 dpi is where OCR accuracy starts falling apart on small type.
If you have no scanner at all, a phone camera plus a scanning app produces a perfectly serviceable PDF for personal filing. The limit shows up with legal or tax documents where a third party may dispute the image quality — there, a flatbed or ADF scan is the safer choice. Receipts for expense claims are the one category where phone scans are routinely accepted.
Step 2: Pick a Scan-to-PDF Tool You Already Own
Four routes cover almost every situation: the scanner maker's bundled software, the phone's built-in scanning feature, a dedicated scanning app, and a multifunction printer's front-panel scan-to-email or scan-to-folder function. Each has a distinct failure mode worth knowing before you commit a hundred pages to it.
The scanner maker's software, such as Epson ScanSmart, Canon IJ Scan Utility, or Brother iPrint&Scan, ships free with the hardware and exposes the full resolution and OCR controls. It's the most reliable option when you need to change settings and the least pleasant to look at. Front-panel scanning on a multifunction printer sends the PDF straight to email or a network folder without a computer in the loop, which is ideal for shared office equipment.
Phones handle this natively now. Apple's Notes app has a built-in document scanner that produces a PDF, and Google Drive's Android app has a scan button that does the same. Both auto-detect page edges and correct perspective, which is what makes phone scanning usable at all. Third-party apps like Adobe Scan and Microsoft Lens add OCR and PDF/A export but increasingly gate features behind subscriptions.
| Tool | Where it lives | OCR included | Best for | Known limitation |
|---|---|---|---|---|
| Epson ScanSmart / Canon IJ / Brother iPrint&Scan | Desktop, bundled | Yes, in scan settings | Full control over dpi and color | Interface varies wildly by model |
| Apple Notes scanner | iPhone, iPad | Live Text on supported devices | Quick one-off pages | Manual sharing to PDF each time |
| Google Drive scan | Android | Yes, via Drive | Filing straight to cloud | Requires Google account |
| Adobe Scan | iOS, Android | Yes, free tier limited | OCR plus PDF/A export | Some features are paid |
| Printer front panel | Multifunction printer | Model-dependent | Shared office, no PC needed | Limited settings on the panel |
Pick the bundled desktop software when you need to change resolution, compression, or OCR language. Pick the phone when the alternative is not scanning at all. Avoid juggling three tools for the same document type — consistent output matters more than the best possible tool for each page.
Step 3: Set Resolution, Color Mode, and File Size Before Scanning
Three settings account for most of the quality-versus-size tradeoff, and each has a defensible default. Resolution controls how much detail the sensor captures. Color mode determines whether the file is one bit per pixel, grayscale, or full color. Compression decides how aggressively the PDF packs that data. Change these once, save the profile, and stop thinking about them.
| Setting | Default to use | When to change it | File size effect |
|---|---|---|---|
| Resolution | 300 dpi | 600 dpi for small print or handwriting | Doubles roughly with each doubling |
| Color mode | Grayscale for text | Color for stamps, signatures, photos | Color files run 3–5× grayscale |
| Compression | JPEG medium for color, CCITT for black-and-white text | Lossless for archival | Higher compression, softer edges |
| Duplex | On if the printer supports it | Off for single-sided originals | Halves page count saved per sheet |
| Auto-crop / deskew | On | Off for already-aligned originals | No size effect, big quality effect |
Color mode is where people waste the most space. A signed letter scanned in full color at 300 dpi might land at 2 MB per page; the same page in grayscale often falls under 400 KB, and in black-and-white with CCITT compression under 100 KB. Grayscale preserves pencil annotations and faded ink that pure black-and-white would drop.
For a 50-page document, that difference matters. At 2 MB per page in color you're producing a 100 MB PDF that many email servers will reject. According to Google's Gmail help documentation, the attachment limit is 25 MB, which means the color-everything approach breaks at roughly page 12. Grayscale at 300 dpi keeps you inside standard email limits for most document lengths.
Step 4: Turn On OCR and Choose the Right Language Pack
Text recognition, usually labelled OCR in scanning software, is the setting that separates a searchable PDF from a photograph of paper. It works by analyzing the image, identifying character shapes, and writing an invisible text layer over the page. Once that layer exists, Ctrl+F finds words, copy-paste pulls sentences, and cloud storage indexes the contents.
Language selection is the setting people miss. If your document is in Spanish and your OCR language is set to English, accented characters and common words will be garbled in the text layer even though the image looks perfect. Scanning software from Epson, Canon, and Brother typically defaults to the operating system's language, not the document's. Set it per job if you regularly scan in more than one language.
Handwriting is the honest limitation. Modern OCR from Tesseract, the open-source engine behind many apps, and from Adobe's engine handles printed text extremely well and cursive only moderately. A form filled in by hand may produce a text layer with 60 to 80 percent accuracy, which is enough for keyword search but not enough to trust for copying data. Signatures and initials often vanish from the text layer entirely.
Check whether OCR worked before you close the file. Open the PDF, press Ctrl+F or Command+F, and search for a distinctive phrase from the middle of the document. If the search highlights the phrase, the text layer is there. If nothing is found, OCR was off, the language was wrong, or the resolution was too low — and the fix is to rescan that job rather than run OCR on a bad image.
Step 5: Save, Name, and File the PDF So It Survives
Naming conventions are unglamorous and decisive. A file called scan001.pdf is functionally lost two weeks after creation. A file named 2026-11-12_contract_acme-signed.pdf sorts chronologically, identifies the counterparty, and flags its status. Use ISO 8601 date format, year first, because it sorts correctly in every file manager and cloud service without extra configuration.
Save location matters next. Local folders are fast and fragile. Cloud storage with version history, such as Google Drive, Dropbox, or OneDrive, gives you rollback if a scan gets corrupted or a bad OCR pass overwrites a good one. For documents you'll need to prove existence of later — contracts, tax filings, insurance policies — keep both a local copy and a cloud copy on separate providers.
Consider PDF/A at save time if the document has legal or archival weight. The option appears in Adobe Acrobat's "Save as Other" dialog, in many scanning apps' export settings, and in free tools like PDF Arranger on Linux. It's the one setting where a little friction now prevents a real problem in a decade. Standard PDF is fine for anything you'd shrug at losing.
If a scan comes out wrong — pages out of order, skewed, or missing the last sheet — rescan rather than repair. Fixing a bad scan in a PDF editor takes longer than running the pages through again, and the repair often leaves artifacts that confuse OCR. ADF misfeed is the common culprit; check the feeder's paper guides and fan the stack before retrying.
Step 6: Fix Common Scanning Failures Before You Rescan
Five failures account for nearly every unusable scan, and each has a specific diagnostic. Working through them in order saves rescanning the same stack repeatedly. Most have nothing to do with the software.
| Symptom | Likely cause | Fix before rescanning |
|---|---|---|
| Pages come out skewed | Feeder guides not snug against stack | Adjust guides, fan the paper |
| Faint or washed-out text | Compression too aggressive, or dirty glass | Clean platen, raise quality setting |
| Dark bands down the page | Dust or toner on the scanner glass | Clean glass with lint-free cloth |
| Text layer missing | OCR off, wrong language, or dpi below 200 | Rescan at 300 dpi with OCR on |
| File too large to email | Color mode, or excessive dpi | Switch to grayscale, cap at 300 dpi |
| Multi-page original splits | Feeder grabbing two sheets | Fan stack, reduce stack height |
The file-too-large failure is the most common one in practice, and it's invisible until someone tries to send it. Check the file size before emailing, not after the bounce. If you're over 25 MB and can't rescan, use a cloud link rather than compressing further — compression that squeezes a 300 dpi grayscale scan below 1 MB per page starts degrading OCR accuracy.
Edge cases exist. Thermal receipts fade, and scanning them at 600 dpi in color may be the only way to capture the ghost of the print before it vanishes. Carbon-copy forms produce layers of text that OCR reads as noise. Passports and ID cards need color and often need both sides on a single sheet, which most scanning software supports under a "combine sides" or "ID card" mode. Recognise these before you set up the job, because rescanning faded paper is a losing race.
How Scanning Documents to PDF Compares With Alternatives
Scanning to PDF isn't always the right output, and knowing when to choose something else prevents rework. The three realistic alternatives are photo formats like JPEG, plain text export, and cloud-first capture flows that skip the local file entirely.
| Output | Searchable | Fidelity to original | Best use | Weakness |
|---|---|---|---|---|
| Scanned PDF with OCR | Yes | High | Contracts, receipts, filings | File size grows with page count |
| JPEG photo | No | High | Single pages, photos | No text layer, no pagination |
| Plain text export | Yes | Low | Notes, drafts | Loses layout, signatures, formatting |
| Cloud capture flow | Yes, if OCR on | High | Receipts, expense claims | Needs connectivity, account |
For a document you'll only look at visually, a JPEG is fine and smaller. For anything you'll search, cite, or file, PDF with OCR is the correct default. Plain text export is useful when you only want the words and not the look of the page — transcribing a printed page to editable text, for example.
The cloud-first option is the one that's quietly replaced local scanning for expense receipts. Point the phone at the receipt, let the app OCR it, and the file lands in the cloud already tagged and searchable. For personal recordkeeping, that's faster than scanning to a local folder. For documents that need to stay off third-party servers, the local scan remains the only compliant route.
Why 300 dpi and OCR Are the Two Settings That Matter Most
If you remember one thing after reading this, make it that resolution and text recognition do most of the work. Everything else is refinement. A 300 dpi scan with OCR produces a searchable, shareable, archivable document. A 1200 dpi scan without OCR produces a huge, unsearchable file. The second is worse despite costing more effort.
The 300 dpi figure isn't arbitrary. It's roughly the resolution at which printed text at 8 to 10 point produces enough pixels for OCR engines to recognize characters reliably. Below 200 dpi, small type starts fusing; above 600 dpi, you're storing detail no human eye or OCR engine will use. The International Organization for Standardization's archival standards point to similar ranges for text documents in PDF/A.
OCR accuracy on printed text from well-lit 300 dpi scans routinely clears 99 percent for standard fonts, based on published Tesseract accuracy benchmarks, though exact numbers depend on page quality and font. That's good enough for search and for most copying. It's not good enough to trust for legal data extraction without a visual check — build that check into your workflow for anything that matters.
Pair the two settings and most of the downstream problems disappear. File size stays reasonable, search works, archival formats are available, and rescanning drops to near zero. The settings are set once per profile and then forgotten, which is the whole point.
Scanning Volume, Time, and Cost in Practice
Scanning isn't free, and the time and storage costs are worth knowing before you commit to a digitisation project. A flatbed scan of one page takes roughly 10 to 20 seconds including placement, depending on resolution. An ADF scanning 40 pages duplex at 300 dpi typically takes 60 to 120 seconds for the whole stack, with far less user attention.
Run the arithmetic on a bad scan. A 200-page project scanned in color at 600 dpi produces roughly 400 MB to 1 GB of files. The same project in grayscale at 300 dpi with OCR typically lands somewhere between 20 and 80 MB. The grayscale version is more useful for search and filing, and it fits in email and cloud storage without further processing.
Device costs vary and shouldn't be quoted as fixed current prices. A dedicated flatbed scanner, a multifunction printer with ADF, and a phone are all valid entry points, and the bundled software that ships with the hardware is usually enough. Cloud storage runs on subscription tiers whose prices change; verify current tiers on the provider's site rather than planning against a remembered figure.
Time computation matters for projects rather than single documents. At 15 seconds per page on a flatbed, a 500-page box takes a bit over two hours of active work. An ADF at 3 seconds per page takes 25 minutes. The bottleneck is usually page preparation, not scanning — removing staples, flattening creases, and checking for misfeeds.
FAQ
Can I scan documents to PDF without a scanner?
Yes. A phone camera plus the built-in scanning feature in Apple Notes or Google Drive produces a valid PDF with auto edge detection and perspective correction. The tradeoff is image quality on glossy or curled paper, where a flatbed remains more reliable. For legal or archival documents, use a scanner if one is available and fall back to a phone only when it isn't.
Why is my scanned PDF not searchable?
The text layer is missing, which almost always means OCR was switched off or the language was set wrong in the scan profile. A secondary cause is scanning below 200 dpi, where character shapes fuse and OCR fails. Search for a word you know is on the page — if nothing highlights, rescan with OCR enabled at 300 dpi in the correct language.
Should I scan at 300 or 600 dpi?
300 dpi for printed text documents, 600 dpi for documents containing small print, handwriting, or fine detail you intend to OCR. Scanning text above 600 dpi adds file size without improving readability. Below 200 dpi you're likely to lose OCR accuracy and should rescan.
What does OCR stand for?
Optical Character Recognition. It's the process that analyzes a scanned image, identifies character shapes, and writes an invisible text layer over the page so the PDF becomes searchable and copyable. The technology is standard across scanner software and phone apps; what varies is accuracy, which depends on scan resolution and document quality.
Is a scanned PDF legally valid?
Usually yes, and in most jurisdictions an electronic signature on a scanned PDF is accepted where the parties have agreed to electronic signatures. For documents that must be presented in their original form, or where a specific jurisdiction requires notarization or wet ink, a scan is a copy rather than a replacement. Check the rule for your jurisdiction and the receiving party's requirements before relying on a scan for anything legally binding.
How do I make a scanned PDF smaller?
Switch from color to grayscale, cap resolution at 300 dpi, and enable auto-crop and deskew. These three changes typically reduce a file to a fifth to a tenth of its original size. If the result is still too large, share via a cloud link rather than compressing further, because aggressive compression degrades the text layer.
Can I scan multiple pages into one PDF?
Yes, and most scanning software supports it by default. On a flatbed, chose the "multi-page" or "add page" option between captures. With an ADF, stack the pages, enable duplex if both sides are printed, and save as a single PDF rather than one file per page. Check the page count in the finished file before filing.
What's the difference between a scanned PDF and a PDF/A scan?
PDF/A is an archival subset that forbids encryption and embedded media and requires all fonts to be embedded, making it suitable for long-term retention. A standard scanned PDF may include compression or features that violate archival rules. Choose PDF/A for deeds, certificates, and permanent records, and standard PDF for everyday filing.