Convert images, HEIC, and PDF pages
Open PNG, JPG, JPEG, HEIC, HEIF, or PDF files up to 20 MB. You can also paste a screenshot, load a direct image URL, or capture text with a camera.
Free private image and PDF to text converter
Online OCR for Images and PDFs
Extract editable text from screenshots, scanned PDFs, receipts, photos, PNG, JPG, HEIC, and HEIF files. Correct the crop, rotation, perspective, and text regions before private in-browser recognition.
Local Tesseract.js OCRPDF & HEIC supportPrecision multi-region cropping
01 / Extract
Convert an image or PDF to editable text
Upload a file, paste a screenshot, enter an image URL, or use your camera. Refine one or more text regions, then review and copy the locally recognized result.
Upload an image
Step 01
Drop your image here
or browse your files
PNG, JPG, HEIC, HEIF, or PDF · up to 20 MB
PNG, JPG, HEIC, HEIF, or PDF · up to 20 MB
Paste an image
Step 01
Use an image from your clipboard
Copy an image or screenshot, then paste it here with Ctrl+V or use the button.
Load an image URL
Step 01
Use a public image link
Enter a direct HTTPS link to a PNG or JPEG. The image host must allow browser access.
Camera scanner
Step 01
Scan text with your camera
Start the camera, hold the text steady in the frame, then capture the image.
Camera access requires HTTPS or localhost. Your camera image stays on this device.
Preparing image
Reading the selected file locally…
Crop image
Drag to crop. Hold Space and drag to pan. Arrow keys nudge by one pixel.
Finding the document edges…
C CropH Hand modeSpace Hold to panR RotateZ ZoomEsc Reset crop↑↓←→ Nudge
Automatic cleanup: text is resized, straightened, contrasted, and thresholded as needed before recognition.
Advanced OCR settings Automatic
Extracted text
Step 02
No text extracted yet
Local processingNothing is uploaded
Smart cleanupSharper text contrast
Editable resultCopy or revise freely
Private OCR with document-grade cropping tools
This free OCR converter prepares difficult scans before recognition and processes the source locally in your browser. Use it for receipts, invoices, labels, forms, ID cards, screenshots, and scanned document pages.
Crop, rotate, deskew, and correct perspective
Use freeform or aspect-locked cropping, zoom and pan, keyboard pixel nudging, edge snapping, grid guides, border trimming, manual rotation, and four-corner perspective correction.
Extract several text regions in one run
Automatically detect text blocks, remove unwanted regions, or add manual crops. Multi-region mode recognizes every retained region independently and combines the text in order.
Improve difficult text automatically
Automatic preprocessing can resize small text, normalize contrast, deskew slight tilts, and retry grayscale, adaptive threshold, or original-color inputs when confidence is low.
Choose the document layout
Recognition presets cover automatic pages, text blocks, sparse text, single lines, single words, receipts and columns, serial numbers, code, and numeric content.
Keep uploaded documents private
Image preparation and OCR run locally with JavaScript and WebAssembly. Uploaded, pasted, camera, HEIC, and PDF content is not sent to a MonkeyTactics application server.
Frequently Asked Questions
-
Is my uploaded image private?
Yes. Uploaded, pasted, and camera images are processed locally and are never sent to MonkeyTactics or stored in a database. When you use an image URL, your browser retrieves it directly from that image host. See our Privacy Policy for more information.
-
What image formats can the OCR utility read?
The utility accepts uploaded PNG, JPG, JPEG, HEIC, HEIF, and PDF files up to 20 MB, plus pasted images and direct HTTPS image URLs whose hosts permit browser access. HEIC, HEIF, and PDF files are converted locally before OCR. Clear screenshots and sharp, evenly lit scans generally produce the best results.
-
How accurate is browser-based OCR?
Accuracy depends on image quality, font clarity, text size, orientation, and contrast. Automatic cleanup can resize small text, correct slight skew, normalize dark backgrounds, and retry adaptive thresholding when confidence is low. Handwritten or highly stylized text may still be difficult to read.
-
Why does the first OCR scan take longer?
The first scan downloads and initializes the Tesseract.js recognition engine and English language data in your browser. Later scans during the same session are usually faster because the worker remains ready.
-
What is Tesseract.js and how does it power this tool?
Tesseract.js is an open-source JavaScript and WebAssembly implementation of the Tesseract optical character recognition engine. It runs the recognition process inside a dedicated browser worker, using trained English language data to identify letters, words, and line structure in the preprocessed image. This allows the OCR Utility to extract text without sending the image to an application server or requiring an API key. Recognition quality still depends on the source image, so clean printed text generally performs better than handwriting, decorative fonts, or blurred photographs. Tesseract.js is provided under the Apache License 2.0, which permits use, modification, and distribution subject to the license terms.
-
How can I improve an inaccurate result?
Crop the image with a little space around the text, trim uniform black or white borders, correct sideways pages or perspective distortion, and choose the closest layout and content presets. For difficult photos, try Adaptive black & white; larger dark text on a light background is typically recognized most reliably.
-
Can OCR extract text from multiple parts of an image?
Yes. Auto-Detect Text Region finds text blocks without an eight-region limit, and each detected region can be edited or removed. Enable Multi-Region Mode to recognize every retained region independently and combine the extracted text in region order.
-
Can I straighten a photographed receipt or document?
Yes. Rotate the source when it is sideways, use Auto Deskew for slight text angles, or enable Perspective Correction and drag the four yellow corner handles onto the document corners. Preview shows the corrected rectangular image before OCR.
