How to Make a Scanned PDF Searchable with OCR on Android

OCR stands for optical character recognition. It looks at the picture of a page, finds the shapes that look like letters and numbers, and works out what they say. The result gets stored in the PDF as a hidden text layer that sits invisibly on top of the original image. The page still looks identical to the scan, but now there is real text underneath that your phone can read.

That hidden layer is what makes a PDF "searchable." Once it exists, you can type a word into your reader's search box and jump straight to it, drag to select a sentence and copy it into an email, or let a screen reader speak the page aloud. Nothing about the look changes; you are adding a capability, not altering the scan.

It helps to know the difference between two things people mix up. A searchable PDF keeps the original page image and adds the text layer behind it, so it looks exactly like your scan and you can find and copy words. Text extraction (sometimes called export to Word or to a document) throws away the image and gives you the raw recognized text in an editable file, which is useful when you want to rewrite the content but loses the original layout. Most of the time you want the first kind, and that is what the methods below produce by default.

OCR can only recognize what it can see clearly, so the quality of the scan decides the quality of the result more than the app does. A few minutes spent capturing a good page saves you from proofreading garbage later.

• Light it evenly. Flat, bright, indirect light beats a single harsh lamp that throws shadows across the page. Avoid your own shadow falling over the document.
• Fill the frame and keep it square. Hold the phone parallel to the page so the text is not skewed, and get close enough that the letters are sharp. Scanner apps auto-detect the page edges and straighten the result, which helps recognition.
• Flatten the paper. Curled or folded pages bend the lines of text and confuse OCR. Press the page flat or weigh down the corners.
• Use a plain background. A dark, uncluttered surface under white paper makes edge detection reliable.

If you scan documents often, it is worth picking a capture app you trust rather than relying on whatever is built in. Our roundup of the best scanner apps for Android compares the options on edge detection, OCR quality, and how aggressively they push subscriptions.

Adobe Scan is the most hands-off option because it runs OCR on its own. You install Adobe Scan: AI PDF Scanner, OCR from Google Play (it now needs Android 10 or later), sign in with a free Adobe account, and point the camera at your document. The app finds the page edges, cleans up the image, and as soon as you save, it processes the text in the background. The PDF it produces is already searchable; there is no separate OCR button to hunt for.

To capture: open Adobe Scan, let it detect the page or tap the shutter, add more pages if you need them, then tap the thumbnail and choose Save PDF. Give it a moment after saving, then open the file and try searching. You can also tap the share or export menu to send it out as a PDF, or export the recognized text to a Word file or to plain JPEG images.

The free version covers normal personal use. Adobe limits how many pages get the full OCR treatment per file on the free tier, and lifting that cap (up to 100 pages) is part of the paid Acrobat subscription, whose price varies by region and current promotion. For everyday scans of a few pages at a time, the free tier handles it. One honest note: Adobe Scan does its recognition in the cloud, so it needs an internet connection to finish processing, and your scans pass through Adobe's servers, which matters if the document is sensitive.

If you have a Google account, you already have a scanner and an OCR engine, no extra install required. The Google Drive app has a built-in scanner, and Drive runs OCR on what you save.

Open the Drive app, tap the + (New) button, choose Scan, and capture your pages. Drive auto-detects edges, lets you adjust the crop and color, and saves the result as a multi-page PDF. Once the file is in Drive, the service indexes its text in the background, usually within a few minutes, so you can find the document by searching for words inside it from Drive's main search bar.

There is a useful second step if you want selectable, copyable text rather than just findability. In Drive, open the file's three-dot menu, choose Open with, and pick Google Docs. Google Docs re-runs OCR and drops the recognized text right into an editable document, with the original image usually placed above it. From there you can copy the text or save a clean copy. Google notes a few limits worth respecting: keep the source file reasonably small and sharp, make sure the page is right-side up, and expect complex layouts like multi-column pages and tables to come out imperfectly. This route is free and convenient, but the Open with Google Docs path gives you text rather than a polished searchable PDF that preserves the original layout, so use Adobe Scan when keeping the look matters.

If you used Microsoft Lens, here is the news you need: Microsoft retired it. Lens entered retired status on January 9, 2026, was removed from the app stores around February 9, 2026, and its scanning stops working around March 9, 2026, with scanning moved into the OneDrive app and Microsoft Copilot. So the old Lens advice you may find online no longer applies, and there is no Lens to download.

The replacement works well. The OneDrive mobile app has a built-in scanner, and through late 2025 and into 2026 Microsoft added OCR for scanned PDFs across the mobile apps, on by default. To scan: open OneDrive, tap the camera or scan icon, capture your pages, and save the PDF to your OneDrive. To make an existing scanned PDF searchable, open that PDF inside the OneDrive app, then tap and hold (long-press) anywhere on the page; when the menu appears, choose Recognize text. OneDrive runs OCR and adds the text layer so you can search and select.

This is the natural choice if you already live in Microsoft 365 or keep your files in OneDrive, since the scan, the OCR, and the storage are all in one place. As with Adobe, recognition happens on Microsoft's servers, so it needs a connection and the document passes through the cloud.

Whichever of the three you pick, the workflow is the same shape: capture a clean image, let OCR add the hidden text layer, then confirm it worked by searching for a word you know is on the page.

Do not assume the OCR succeeded just because the app said "done." Take ten seconds to check, because a silently bad recognition is worse than none at all.

The quick test: open the finished PDF in any reader, search for a distinctive word you can see on the page, and try to select a line of text by dragging. If the search finds it and the selection highlights the actual words, the text layer is there and correct. If the search comes up empty or your finger selects nothing, OCR either did not run or did not stick, so re-run it or try one of the other methods.

When recognition ran but produced wrong text, the cause is almost always the source image. Rescan the page with better light and a flatter, squarer shot. If a single word or two came out wrong, you can correct them: in Google Docs the text is directly editable, and in a full PDF editor you can fix the recognized layer. For tidying up scanned PDFs after OCR, reordering pages, or fixing recognized text, our guide to the best PDF editor apps for Android covers apps that let you edit the text layer directly.

OCR is good, not perfect, and pretending otherwise will burn you on the documents that matter most. Recognition accuracy drops sharply in a few predictable situations, and you should expect to proofread anything important.

• Handwriting. Most phone OCR is built for printed type. Cursive and casual handwriting are recognized poorly or not at all. A few specialized tools attempt handwriting, but treat any result as a rough draft to verify, not a transcript to trust.
• Poor lighting and low resolution. Shadows, glare, blur, and a faded or low-contrast original all starve the recognizer of the detail it needs. This is the single most common reason OCR returns gibberish, and it is the easiest to fix by rescanning.
• Unusual fonts and decorative type. Heavily stylized fonts, condensed type, very small print, and old or ornate typefaces trip up recognition. Standard book and document fonts work best.
• Complex layouts. Multi-column pages, tables, and forms often come out with the reading order scrambled, even when each individual word is recognized correctly.
• Mixed or non-Latin scripts. Accuracy depends on the language pack, and pages mixing scripts can confuse the engine.

The rule to carry away: always proofread the recognized text before you rely on it, especially for anything legal, financial, or medical. Search the document for a couple of key figures, like a total, a date, or a name, and confirm they match the page. OCR saves you retyping; it does not replace reading.

Pick the path that matches what you already use and how much the document matters:

• You scan a lot and want it automatic: use Adobe Scan. OCR runs on save, the searchable PDF preserves the original look, and export to Word is there when you need editable text.
• You live in Google's apps: use Drive's built-in scanner for findable PDFs, and the Open with Google Docs trick when you need to copy or edit the text.
• You use Microsoft 365 or OneDrive: scan in OneDrive and long-press, then Recognize text on any existing scan. This is also the answer for former Lens users, since Lens is gone.
• The text matters more than the layout: after OCR, you may prefer to drop it into a notes app for editing and searching alongside your other notes. The best notes apps for Android roundup covers options that store text well.

Whatever the document, treat OCR as a strong first pass that still needs your eyes on the result. If you want to streamline the whole capture-to-search routine on your phone, the wider productivity apps section has more on building a paperless workflow that actually holds up.