Image to Text (OCR) - Free Online Tool

OCR (optical character recognition) extracts text from an image so you can copy, search, or reuse it. Use it to grab text from screenshots, photos of documents, receipts, labels, or error messages when you do not have the original editable file. OCR runs locally in your browser, so your image is not uploaded for processing.

1. Upload an image: Choose a screenshot or photo that contains text.
2. Set options: Pick a minimum confidence level and choose plain text or JSON output.
3. Run OCR and copy: Start recognition, then copy the extracted text.

• Copy text instantly: Turn screenshots and photos into editable text without retyping.
• Filter noisy results: Use minimum confidence to reduce random characters in the output.
• Privacy-first scanning: Recognition happens on your device in the browser.

OCR is never perfect. These options help you trade speed, readability, and precision:

• Minimum confidence: Lower values include more text (but can add noise). Higher values produce cleaner output but may drop faint or stylized text.
• Plain text vs JSON: Plain text is easiest to copy into docs. JSON is useful when you want confidence scores and rough bounding boxes for each detected chunk.
• Join lines: Choose new lines to preserve layout, or spaces when you want one flowing paragraph.

Use case 1: Copy error messages from a screenshot

Paste OCR output into a bug report, support ticket, or search query without manually retyping long stack traces or dialog text.

Use case 2: Digitize receipts and notes

Extract line items, totals, or short notes from a photo so you can paste them into spreadsheets, emails, or expense tools.

Use case 3: Pull text from images in PDFs

If a PDF is a scanned image (not selectable text), take a screenshot of the relevant area and run OCR to make it copyable.

Basic example

Input: Screenshot that says: 403 Forbidden - Invalid API key
Output (plain text):

403 Forbidden - Invalid API key

Advanced example (JSON output with confidence)

Input: Photo of a shipping label
Settings: Output format: JSON; Min confidence: 60%
Output (example):

[
  { "text": "TRACKING", "confidence": 0.98, "boundingBox": { "x": 48, "y": 22, "width": 120, "height": 18 } },
  { "text": "1Z999AA10123456784", "confidence": 0.94, "boundingBox": { "x": 52, "y": 52, "width": 220, "height": 20 } }
]

Advanced example (receipt-style plain text)

Input: Photo of a store receipt
Settings: Output format: Plain text; Join lines: New lines; Min confidence: 50%
Output (example):

Coffee Shop
Latte 4.50
Tax 0.41
Total 4.91

First run takes longer

On first use, your browser loads OCR model files. After that, the models are cached for faster scans.

Output has missing characters or random symbols

OCR accuracy depends on image quality. Blurry photos, low contrast, and skewed angles reduce accuracy. Try a sharper image, crop closer to the text, and increase lighting.

Output is empty

If minimum confidence is too high, most detections can be filtered out. Lower the confidence threshold and try again.

Line breaks look wrong

Some images do not have clean line spacing or alignment, so OCR may join or split lines oddly. Try switching Join lines between new lines and spaces, or crop closer to the text region to reduce layout ambiguity.

• Crop to the text: Smaller regions with clearer text improve accuracy and reduce noise.
• Keep text horizontal: Straighten the image if the text is rotated; OCR works best when lines are level.
• Expect lower accuracy for handwriting: OCR works best on printed text. Handwritten notes often need manual cleanup.
• Clean the result for reuse: After OCR, remove extra whitespace with the whitespace remover or redact sensitive fields with the privacy scrubber.

• Remove formatting and invisible characters after OCR with remove text formatting.
• Count words and characters in extracted text with the online sentence counter.

OCR runs locally in your browser. Your images are not uploaded to a server for processing. If you are scanning sensitive documents, avoid sharing screenshots and close the tab after copying the text.