OCR and document extraction tools

Short answer: To extract text or tables from scans, screenshots and photos, choose OCR for text or image-to-Excel for structured tables, then review the result. QuickMerge processes the source file in your browser without uploading it.

These tools turn pictures of text back into text you can edit, search or paste into a spreadsheet. Recognition runs in your browser using Tesseract.js, so the image itself is not uploaded to QuickMerge — though the OCR engine does download its own language data the first time you use it.

4 tools in this canonical category. Supported operations use local browser processing; see how to inspect a specific workflow.

Get text out of an image or scan

Use these when the words you need are locked inside a picture.

  • OCR — Extract readable text from a scanned PDF, photo or screenshot.
  • PDF summarizer — Reduce a long document to its key sentences on your device.

Get tables into a spreadsheet

For data that is trapped in a screenshot, a scan or a block of raw text.

  • Image to Excel — Read a table out of a screenshot or photo and download it as XLSX or CSV.
  • Text to Excel — Turn pasted text, TXT or CSV data into a real spreadsheet file.

Pick the tool by what you need out, not what you put in

All of these read a document and give you back its content, but the useful distinction is the shape of the output. If you want prose — a letter, a report, a contract — OCR returns a stream of text. If you want a table with rows and columns intact, image to Excel is built for that, and plain OCR will disappoint you because it flattens the structure.

If the file already has a text layer — a PDF produced by software rather than a scanner — you do not need recognition at all. PDF to Word exports it directly and keeps far more formatting than OCR ever could. Running OCR over a text-based PDF is a common mistake that produces worse output than doing nothing.

What decides whether recognition works

Resolution first: below roughly 300 DPI, character shapes stop being distinct enough to separate reliably. Orientation second, and it is the easiest to fix — straighten the pages before recognising, because skewed text is the single most common cause of poor results. Contrast third: dark text on a clean light background reads best, and shadows across a photographed page hurt more than most people expect.

Handwriting is out of scope. These engines are trained on printed text, so cursive and handwritten notes will not produce usable output regardless of scan quality.

Guides for these tools

Questions people ask

How accurate is the OCR?

Accuracy depends on the image. Clean screenshots and straight, high-contrast scans of printed text give the best results. Photos taken at an angle, low-resolution images, glare, handwriting and very small type all reduce accuracy. Every tool shows you the extracted text or table so you can correct it before you download.

Does OCR work on handwriting?

The engine is trained on printed text. Handwriting is generally not recognised reliably and usually needs manual correction.

Is the image uploaded to QuickMerge?

No. The image is read in your browser. The OCR engine does fetch its own JavaScript and language data from a public CDN the first time it runs, but your image is not part of that request.

How do I turn a scanned PDF into an editable document?

Run OCR to recover the text, then save that text where you need it. If the PDF was never scanned and already has a text layer, skip recognition entirely and use PDF to Word, which preserves far more of the original formatting.

Why is my OCR output full of errors?

In order of likelihood: the scan is below about 300 DPI, the page is rotated or skewed, contrast is poor, or the wrong language is selected. Resolution and orientation account for most bad results, and both are fixable before you try again.

Can these tools read handwriting?

No. Recognition here targets printed text. Handwritten and cursive material will not produce usable output.

How do I get a table out of a screenshot?

Use Image to Excel, which reconstructs rows and columns. Plain OCR returns the characters but loses the table structure, which is rarely what you want from a table.

Is my document uploaded for recognition?

No. Recognition runs in your browser, which is the point when the document is an invoice, a medical letter or an identity record.

Other categories