How to Extract Text from an Image (OCR Guide)

Knowing how to extract text from an image is one of the most useful document skills you can have. Whether you are looking at a scanned contract, a screenshot of an email, or a photo of a whiteboard, OCR (Optical Character Recognition) technology can read the pixels and hand you back editable text in seconds.

Key takeaways

• OCR converts image pixels into selectable, searchable characters.
• A browser-based tool handles most everyday tasks without any software to install.
• Image quality directly affects how accurate the extracted text will be.
• Scanned PDFs need OCR just as much as plain images do.
• Cleaning up a poor scan before running OCR produces noticeably better results.
• Free plan users can process up to 10 OCR pages per day; Premium supports up to 200 pages per job.

What OCR actually does

OCR software analyzes the shapes of characters in an image and maps them to letters, numbers, and symbols. Modern engines handle dozens of languages, mixed fonts, and even handwriting in some cases. The output is plain text or a searchable document layer you can copy and paste anywhere.

The three things that affect accuracy most are resolution, contrast, and skew. A crisp, straight scan at 150 DPI or higher will almost always produce clean output. A blurry phone photo taken at an angle is harder for any engine to parse.

When a browser tool is enough

For everyday tasks — pulling a quote from a screenshot, copying a table from a scanned page, or grabbing an address from a photo — a browser tool covers everything without installation or account setup.

Use a browser tool when:

• The image is under 50 MB (free plan) or 500 MB (Premium plan).
• You have a handful of pages rather than hundreds.
• You need results quickly without configuring software.
• You are on a shared or work computer where installing software is not an option.

When desktop software makes more sense

Desktop OCR applications become worthwhile when you are processing hundreds of documents in bulk, need to build automated pipelines, or must keep files entirely offline for compliance reasons. They also let you batch-rename output files and integrate with local databases.

For most people reading this guide, a browser tool handles everything they need.

How to extract text from an image: step by step

Step 1 — Prepare your image

Before uploading, check these three things:

1. Resolution: Aim for at least 150 DPI. If you are photographing a printed page, hold the camera steady and use good lighting.
2. Straighten the page: A skewed image causes OCR engines to misread whole lines. Most phone cameras have a grid overlay — use it.
3. Crop out clutter: Remove table edges, fingers, or background noise that surrounds the text area. This keeps the engine focused on the content you care about.

Step 2 — Upload to an OCR tool

1. Open Image to Text in your browser.
2. Click Choose file or drag your image (JPG, PNG, TIFF, BMP, and similar formats are all accepted).
3. Select the language of the text in the image if the tool offers that option — choosing the correct language improves character matching.
4. Click Extract text or the equivalent action button.
5. Wait for processing. Most images finish in a few seconds.

Guest users can run 1 job per day with files up to 25 MB. Free accounts get 2 jobs per day with files up to 50 MB and up to 10 OCR pages. Premium users can handle files up to 500 MB and 200 OCR pages per job.

Step 3 — Review and copy the output

Once the tool returns results, scan through the text for obvious errors before copying it. Common OCR mistakes include:

• The letter l confused with the number 1 or capital I.
• Broken words at line ends if the original had hyphenation.
• Special characters like em dashes replaced with hyphens.

Select all the text, copy it, and paste it into your word processor or notes app. Run a quick spell-check there to catch any remaining recognition errors.

Step 4 — Extract text from a scanned PDF

Scanned PDFs are images dressed up as PDF files. There is no real text layer inside — just pixel data. To make that text selectable and searchable, you need to run OCR against the PDF specifically.

1. Open OCR PDF.
2. Upload your scanned PDF.
3. Choose your language and output preference (searchable PDF or plain text).
4. Download the result. You can now select, copy, and search any text in the document.

If the scanned pages look faded, skewed, or speckled, run them through Clean Up Scanned PDF first. Deskewing and despeckling before OCR can make a noticeable difference in how cleanly the text comes out.

Step 5 — Convert a clean PDF to plain text

If your PDF already has a real text layer (meaning it was created digitally, not scanned), you do not need OCR at all. Use PDF to Text to pull out all the text content directly, which is faster and perfectly accurate.

Not sure which type you have? Try selecting a word in your PDF viewer. If you can highlight it, the text layer exists. If nothing highlights, it is a scanned image and needs OCR.

Step 6 — Turn a physical document into a searchable PDF

If you are starting from paper rather than a digital file:

1. Scan the document using your printer or a phone scanning app. Save as PDF or JPG.
2. If you saved as JPG, use Scan to PDF to bundle pages into a single PDF.
3. Run Clean Up Scanned PDF to fix rotation, brightness, and noise.
4. Apply OCR PDF to add a searchable text layer.
5. Download your finished, searchable document.

Tips for better OCR results

• Higher resolution beats lower resolution. A 300 DPI scan consistently outperforms a 72 DPI screenshot.
• Black text on white background is the easiest input for any OCR engine. Colored backgrounds and light-colored text cause more errors.
• Avoid heavy compression. JPEG compression introduces artifacts around letter edges that confuse character recognition. If you can choose between PNG and JPEG for a screenshot, PNG preserves edges better.
• Handwriting is harder. Printed text recognition is mature and reliable. Handwriting recognition is improving but still less consistent, especially with cursive or irregular lettering.
• Multi-column layouts sometimes confuse the reading order. Check that columns were not merged into a single jumbled paragraph in the output.

Next steps

Once you have extracted the text you need, you may want to turn it back into a document. If your output is plain text, Text to PDF can wrap it into a shareable PDF in one step. If you ended up with a Word document, Word to PDF handles the final conversion. The workflow from image to polished document is shorter than most people expect.