Build a Searchable Archive

A scan is a picture of a page. A box of pictures cannot be searched, cannot have its text selected, and cannot be indexed — until each page carries a text layer. The archive job adds that layer: clean the captures, recognize the text on-device, and store the result in the format built for long-term preservation.

The Job, in Order

1. Clean up the captures. Phone photos and old scans arrive crooked, grey, and uneven. enhance a scanned document — in your browser, free — straightens the pages and evens the contrast. This is worth doing first: the recognition step reads a clean page far better than a messy one.

2. Recognize the text on the device. OCR a PDF — on-device recognition, free, no upload — reads each page and adds the text layer, in place, with nothing transmitted. This is the step that turns a picture into a document.

3. Store the archive in the archival format. convert a PDF to PDF/A — in your browser, free — writes the result as PDF/A, the ISO standard for long-term preservation, which is the format an archive is supposed to be stored in. Where the result has to be validated against the standard, validate PDF/A compliance runs as a cloud step with a free account’s daily and monthly quota.

Verify the Archive

Search for a word that must be in the archive and confirm it is found. Select a line of recovered text. Copy a sentence and compare it against the page. Recognition is probabilistic, so verification is the job, not an extra — and it is the only way to know which pages need a second capture instead of a second guess.

When the Archive Is Also Huge

Scans are image files first, so a searchable archive is often a heavy one. Once the text layer is in place, the size question is a separate job: the scan-compression guide walks the size-versus-quality trade without destroying the legibility you just built, and compress an image — in your browser, free — is the lever that actually shrinks a scan.

Frequently Asked Questions

Will it read handwriting? No, and no general-purpose OCR tool honestly claims it will. Printed and typed text is what this step reliably recovers; handwritten pages stay pictures.

The recovered text has errors — why? Recognition follows the input: a skewed, low-contrast, or blurred page gives a text layer with gaps. Clean the capture first, and when a page still comes back wrong, the fix is usually a better original, not another pass.

The archive got bigger after OCR — is that normal? Yes: the text layer is added alongside the images, so the file grows by the size of the words. Compress after recognizing, not before, so you never throw away the pixels the text was recovered from.

---

A searchable archive is three steps done in the right order: clean the pages, recognize the text on the device, and store the result as PDF/A. Do it that way and the pile becomes something you can search ten years from now — while the scans themselves never left your machine.

Do It Directly

Ready to build the archive in front of you? enhance a scanned document the captures first, then OCR a PDF — on-device recognition, free, no upload — to add the text layer, and convert a PDF to PDF/A for the long-term copies. Heavy result? The scan compression guide shows how to shrink it without blurring the words you just recovered.

Related reading

• Make a Scanned PDF Searchable — the single-file version of the same job
• How to Compress a Scanned PDF and Keep the Text Crisp — shrinking an image-heavy archive without destroying legibility
• Local vs Cloud Processing — which step of this archive job belongs in the browser and which can use the cloud