How to turn photos and scans into one searchable PDF

Turn photos or scans into one searchable PDF in three steps with OnCloudPDF: Images to PDF makes a PDF from your pictures, Merge PDF joins it with other PDFs if needed, and OCR PDF adds an invisible text layer in one of eight languages. OCR runs in your browser, can be slow on long files, and its accuracy depends on how clear the scans are.

Updated October 29, 2026 · Runs in your browser, no upload, no account

What you are building and why it helps

A photo of a page is only a picture: you cannot search it, select words or copy a paragraph. A searchable PDF keeps the picture of the page but adds an invisible layer of recognized text on top, so Ctrl+F finds words and you can copy text. The step that creates that layer is OCR, optical character recognition.

A typical case is a stack of receipts, a signed letter photographed on a phone, or a few pages from a book. You have several image files, possibly a scanner PDF as well, and you want one document that is easy to store and search. The workflow below does that without installing software, and the files stay in your browser.

Step 1: take or collect good images

OCR accuracy starts at capture. Tesseract, the engine OCR PDF uses, recognizes clean, high-resolution scans in the right language best; blurry, skewed or low-contrast images produce more errors. Practical habits help more than any setting:

• Photograph the page flat, from directly above, with even light and no shadow across the text.
• Fill the frame with the page so the letters are large and sharp.
• Keep the page straight. The tools do not deskew, so a tilted page stays tilted.
• Name the files so they sort in reading order, such as page-01.jpg and page-02.jpg.

Step 2: Images to PDF

Select all the images at once in Images to PDF. Accepted types are JPG/JPEG, PNG and GIF. Each image becomes one page whose size equals the image's pixel dimensions, and pages follow the order in which the files were selected. There is no page-size or fit-to-page option, and a GIF is redrawn as a JPEG using only its first frame.

Two practical points. First, the order: there is no reorder step in this tool, so if a page lands in the wrong place, fix it afterwards with Reorder Pages on the finished PDF. Second, size: JPEG and PNG files are embedded as they are, so the PDF is roughly as large as the sum of your images. Phone photos are large, so a 20-page document can be big. Compress only optimizes the structure and will not recompress these images, so if the size matters, reduce your photos before this step.

Download the result, named images_combined.pdf.

Step 3: Merge, only if you have more than one PDF

If you also have a scanner PDF, or you made several batches, use Merge PDF. Select or drop the files, drag the previews into order, click Merge and download OnCloudPDF_merge.pdf. All pages of every file are merged and they are copied unchanged. If you need only some pages of one file, use Extract Pages or Delete Pages first. An encrypted PDF is not unlocked by Merge.

Do the merge before OCR rather than after. Recognition then runs once on a single file, and the final document has one consistent text layer.

Step 4: run OCR

Open OCR PDF, select the combined file, choose the document language and click Run OCR. The list contains English, Spanish, French, German, Italian, Portuguese, Chinese (Simplified) and Japanese, and you pick one language per run. A document in two languages is recognized with only the one you choose, so words in the other may come out wrong.

Each page is rendered at 1.5x scale and recognized by Tesseract.js in your browser; the output pages are rebuilt from those rendered images with an invisible, positioned text layer. Because of that, the original text and vector content of the PDF is not preserved: even a PDF that already had real text becomes a picture plus an OCR layer. Use OCR only on pages that really are images.

Three practical limits to know before you start. The first run downloads the Tesseract engine and language data (not your file) from a CDN, so it needs an internet connection; later runs reuse the cached data. Processing is page by page on your device and can be slow for long documents, so keep the tab open and expect to wait. And there is no page or size limit enforced, so very large files are bounded by your browser's memory.

When it finishes, check the text preview and then test the PDF: search for a word you know is on a page, and copy a paragraph into a text editor. Download the searchable PDF.

A note on PDF to Scan

PDF to Scan goes in the opposite direction. It takes a normal PDF and makes it look scanned: each page is rendered as an image, light grain and a faint paper tint are added, and the result is saved as a ZIP. It does not run OCR and does not add a text layer, so its output is not searchable. Do not run it before OCR on a document you want to search; if you want both a scanned look and a text layer, run PDF to Scan first and OCR PDF on the result.

Summary of the workflow

When another tool is a better choice

OCR PDF is meant for occasional documents. If you need a language that is not in the list, very high accuracy on poor scans, tables recognized as tables or batches of hundreds of pages, desktop OCR software or a server-based service will usually do better, with the privacy trade-off explained in Free PDF tools compared. To confirm that your PDF now has real text, Word Count reports 0 words for a PDF without a text layer and a real number once OCR has worked.

Frequently asked questions