Need to get the words out of a photo? Whether it's a scanned receipt, a whiteboard snapshot, a screenshot of a document, or a photo of a sign, OCR (Optical Character Recognition) turns image pixels into actual text you can copy, edit, and search. This guide explains how to use the free Image to Text tool on this site. The recognition runs on our server: your file is sent over an encrypted connection and deleted as soon as the text has been extracted.
What Is OCR?
OCR stands for Optical Character Recognition. It's the technology that analyses an image and identifies the shapes of letters, digits, and punctuation to produce machine-readable text.
Modern OCR engines use neural networks trained on millions of printed and handwritten characters. They can recognise text in a wide range of fonts, sizes, languages, and image conditions — even at moderate angles or with some blur.
OCR is how:
- Banks digitise cheques and identity documents
- Google Search indexes text inside images
- Postal services read handwritten addresses on envelopes
- Archivists make millions of scanned books searchable
The same technology is available free here.
How to Extract Text from an Image — Step by Step
The Image to Text tool runs Tesseract OCR on our server. Your image is uploaded over an encrypted HTTPS connection, the text is extracted, and the uploaded file is deleted from the server afterwards.
Step 1 — Open the Image to Text tool
Go to want2convert.com/image-to-text. You'll see a simple upload zone and a language selector.
Step 2 — Upload your image
Drag and drop your image into the upload zone or click to browse your device. Supported formats are JPG, PNG, TIFF, BMP, WebP, and PDF, up to 50 MB. For scanned documents, use the highest resolution scan available — 300 DPI or higher produces the best results.
Step 3 — Select the language (if needed)
By default, the engine recognises English. The tool also supports Swahili (Kiswahili), French, and Arabic — choose the language of the text in your image from the language dropdown before extracting. Using the correct language model significantly improves accuracy for accented characters and language-specific letter patterns. Other languages are not currently supported.
Step 4 — Click "Extract Text"
Click the button. Your file is sent to our server, where it is enlarged and converted to greyscale to improve recognition and then read by Tesseract. Processing usually takes a few seconds, depending on image size and text density; requests time out after about two minutes.
Step 5 — Copy or download the result
The extracted text appears in a text area. You can:
- Copy to Clipboard — paste directly into Word, Google Docs, or any text editor
- Download .txt — save the result as a
.txtfile
What Makes a Good Source Image for OCR
OCR accuracy depends heavily on the quality of the input image. Here's what makes the biggest difference:
Resolution: Text needs enough pixels to be legible. Mobile phone photos of documents taken in good light at reasonable distance (30–50 cm) typically work well. The minimum practical resolution for clear small print is around 150–200 DPI; 300 DPI is the OCR standard for archival scanning.
Contrast: Dark text on a light background gives the engine the clearest signal. Faded ink, light grey text on white paper, or coloured text on a patterned background all reduce accuracy.
Sharpness: Blurry images — from camera shake or focus problems — are the single biggest cause of OCR errors. Use a stable surface when photographing documents. Phone macro modes work well for small-print documents.
Orientation: Text that is severely tilted or rotated reduces accuracy. The OCR engine can handle minor skew (a few degrees), but pages photographed at a steep angle perform worse. Use the Rotate Image tool to turn your photo the right way up first if needed.
Noise and shadows: Creases, shadows across the page, and background textures reduce accuracy. Photograph documents flat on a light-coloured desk with even lighting.
Accuracy: What OCR Can and Can't Do
OCR is highly accurate on clean, high-contrast, printed documents — often 99%+ character accuracy on good input. The technology has real limits to understand though:
| Input type | Expected accuracy | |------------|------------------| | Clean scanned printed document (300 DPI) | 98–99%+ | | Sharp phone photo of printed text | 90–97% | | Faded or low-contrast print | 70–85% | | Handwriting (neat, consistent) | 60–80% | | Handwriting (cursive, varied) | 40–65% | | Heavily stylised fonts or logos | 30–60% | | Text in an image with heavy noise/background | 50–75% |
For printed text in good conditions, expect near-perfect results. Always proofread the output — a single character error in a number or name can have real consequences.
Common Uses for Image to Text
Scanned Documents and PDFs
Offices still deal with mountains of paper. When you receive a scanned contract, form, or letter as a PDF or image, OCR converts it into a searchable, editable document. After extracting the text, paste it into Word or Google Docs for editing, or save it as a file you can search with Ctrl+F.
Receipts and Invoices
Photographing receipts is a common workflow for expense tracking. OCR extracts the line items, totals, and vendor names so you can paste them directly into a spreadsheet or expense report without retyping.
Business Cards
A quick phone photo of a business card, run through OCR, gives you the contact's name, title, phone number, email, and company — ready to copy into your contacts app.
Screenshots of Articles and Web Pages
If you need to quote from a screenshot or copy text from an image-based PDF (a scanned book, a non-selectable PDF), OCR extracts it as actual text you can use.
Whiteboard Photos
After a meeting, photographs of whiteboard notes often sit in a camera roll never acted upon. OCR turns them into text files or searchable notes.
Book and Archive Digitisation
Researchers and genealogists scanning old books, newspapers, and historical records use OCR to make those archives searchable. Tesseract, the open-source engine used here, is widely used in archival digitisation projects.
Translation Workflow
Extract text from an image first, then paste it into a translation tool. This is far faster than manually retyping text from a photo of a foreign-language document, menu, or sign.
OCR vs Copy-Paste: When to Use Each
If you're working with a digital PDF where the text layer is present, Ctrl+A + Ctrl+C (select all, copy) is faster and more accurate than OCR. Use OCR when:
- The PDF is a scanned image (no text layer)
- The source is a photo or screenshot
- The document doesn't allow text selection (locked PDFs — though you might also want to unlock it first)
- The text is in a physical document, sign, or whiteboard
Frequently Asked Questions
Is the OCR tool really free?
Yes, completely free. No account, no subscription, no watermark on the output.
Are my images safe?
Your image or PDF is sent to our server over an encrypted (HTTPS) connection, processed there with Tesseract OCR, and deleted as soon as the text has been extracted. We don't keep copies and no account is needed. Because the file is processed on a server, avoid uploading documents you are not comfortable sending over the internet.
What languages does it support?
English is the default. You can also choose Swahili (Kiswahili), French, or Arabic from the language selector. Other languages are not currently supported.
How accurate is it?
For clean printed text at adequate resolution, accuracy is typically very high. For handwriting or poor-quality images, accuracy is lower. Always proofread the output for critical documents.
Can it read handwriting?
OCR has limited handwriting recognition. It works best with printed text, and handwritten notes may produce poor results, particularly cursive script.
What's the maximum image size?
Uploads are limited to 50 MB, and requests time out after about two minutes. Very large images can be slow; resizing to 2000–3000 pixels wide with the Resize Image tool before OCR often speeds things up without reducing accuracy.
Can I extract text from a PDF?
Yes, you can upload a PDF directly; its pages are rendered and read one by one. If a PDF is large or you want more control, you can also save pages as JPG or PNG first using the PDF to JPG tool and run OCR on the images.
Why are some characters wrong?
OCR errors most often come from low resolution, blur, low contrast, or unusual fonts. Try improving the input image quality and re-running. For systematic errors (e.g. l being read as 1), manually correct the output after extraction.
Related Tools
- Resize Image — scale up a small image before OCR to improve accuracy
- Rotate Image — straighten a tilted photo before OCR
- PDF to JPG — convert scanned PDF pages to images for OCR
- Compress Image — reduce image size without losing the resolution that OCR needs
- JPG to PNG — convert to PNG for lossless quality before OCR on fine-print documents
- Crop Image — isolate the text area of a photo before running OCR
Try it free — no signup needed
Most tools run in your browser, so your files stay on your device. A few upload the file to our server and delete it afterwards.