OCR any JPEG

Drop a photo, scan, or PDF (up to 2.5GB). We extract the text right in your browser — free, unlimited, and your files never leave your device.

Private and secure

Everything happens in your browser. Your files never touch our servers.

Blazing fast

No uploading, no waiting. Convert the moment you drop a file.

Actually free

No account required. No hidden costs. No file size tricks.

Optical Character Recognition (OCR) turns images of text—scans, smartphone photos, PDFs—into machine-readable strings and, increasingly, structured data. Modern OCR is a pipeline that cleans an image, finds text, reads it, and exports rich metadata so downstream systems can search, index, or extract fields. Two widely used output standards are hOCR, an HTML microformat for text and layout, and ALTO XML, a library/archives-oriented schema; both preserve positions, reading order, and other layout cues and are supported by popular engines like Tesseract.

A quick tour of the pipeline

Preprocessing. OCR quality starts with image cleanup: grayscale conversion, denoising, thresholding (binarization), and deskewing. Canonical OpenCV tutorials cover global, adaptive and Otsu thresholding—staples for documents with nonuniform lighting or bimodal histograms. When illumination varies within a page (think phone snaps), adaptive methods often outperform a single global threshold; Otsu automatically picks a threshold by analyzing the histogram. Tilt correction is equally important: Hough-based deskewing (Hough Line Transform) paired with Otsu binarization is a common and effective recipe in production preprocessing pipelines.

Detection vs. recognition. OCR is typically split into text detection (where is the text?) and text recognition (what does it say?). In natural scenes and many scans, fully convolutional detectors like EAST efficiently predict word- or line-level quadrilaterals without heavy proposal stages and are implemented in common toolkits (e.g., OpenCV’s text detection tutorial). On complex pages (newspapers, forms, books), segmentation of lines/regions and reading order inference matter:Kraken implements traditional zone/line segmentation and neural baseline segmentation, with explicit support for different scripts and directions (LTR/RTL/vertical).

Recognition models. The classic open-source workhorse Tesseract (open-sourced by Google, with roots at HP) evolved from a character classifier into an LSTM-based sequence recognizer and can emit searchable PDFs, hOCR/ALTO-friendly outputs, and more from the CLI. Modern recognizers rely on sequence modeling without pre-segmented characters. Connectionist Temporal Classification (CTC) remains foundational, learning alignments between input feature sequences and output label strings; it’s widely used in handwriting and scene-text pipelines.

In the last few years, Transformers reshaped OCR. TrOCR uses a vision Transformer encoder plus a text Transformer decoder, trained on large synthetic corpora then fine-tuned on real data, with strong performance across printed, handwritten and scene-text benchmarks (see also Hugging Face docs). In parallel, some systems sidestep OCR for downstream understanding: Donut (Document Understanding Transformer) is an OCR-free encoder-decoder that directly outputs structured answers (like key-value JSON) from document images (repo, model card), avoiding error accumulation when a separate OCR step feeds an IE system.

Engines and libraries

If you want batteries-included text reading across many scripts, EasyOCR offers a simple API with 80+ language models, returning boxes, text, and confidences—handy for prototypes and non-Latin scripts. For historical documents, Kraken shines with baseline segmentation and script-aware reading order; for flexible line-level training, Calamari builds on the Ocropy lineage (Ocropy) with (multi-)LSTM+CTC recognizers and a CLI for fine-tuning custom models.

Datasets and benchmarks

Generalization hinges on data. For handwriting, the IAM Handwriting Database provides writer-diverse English sentences for training and evaluation; it’s a long-standing reference set for line and word recognition. For scene text, COCO-Text layered extensive annotations over MS-COCO, with labels for printed/handwritten, legible/illegible, script, and full transcriptions (see also the original project page). The field also relies heavily on synthetic pretraining: SynthText in the Wild renders text into photographs with realistic geometry and lighting, providing huge volumes of data to pretrain detectors and recognizers (reference code & data).

Competitions under ICDAR’s Robust Reading umbrella keep evaluation grounded. Recent tasks emphasize end-to-end detection/reading and include linking words into phrases, with official code reporting precision/recall/F-score, intersection-over-union (IoU), and character-level edit-distance metrics—mirroring what practitioners should track.

Output formats and downstream use

OCR rarely ends at plain text. Archives and digital libraries prefer ALTO XML because it encodes the physical layout (blocks/lines/words with coordinates) alongside content, and it pairs well with METS packaging. The hOCR microformat, by contrast, embeds the same idea into HTML/CSS using classes like ocr_line and ocrx_word, making it easy to display, edit, and transform with web tooling. Tesseract exposes both—e.g., generating hOCR or searchable PDFs directly from the CLI (PDF output guide); Python wrappers like pytesseract add convenience. Converters exist to translate between hOCR and ALTO when repositories have fixed ingestion standards—see this curated list of OCR file-format tools.

Practical guidance

  • Start with data & cleanliness. If your images are phone photos or mixed-quality scans, invest in thresholding (adaptive & Otsu) and deskew (Hough) before any model tuning. You’ll often gain more from a robust preprocessing recipe than from swapping recognizers.
  • Choose the right detector. For scanned pages with regular columns, a page segmenter (zones → lines) may suffice; for natural images, single-shot detectors like EAST are strong baselines and plug into many toolkits (OpenCV example).
  • Pick a recognizer that matches your text. For printed Latin, Tesseract (LSTM/OEM) is sturdy and fast; for multi-script or quick prototypes, EasyOCR is productive; for handwriting or historical typefaces, consider Kraken or Calamari and plan to fine-tune. If you need tight coupling to document understanding (key-value extraction, VQA), evaluate TrOCR (OCR) versus Donut (OCR-free) on your schema—Donut may remove a whole integration step.
  • Measure what matters. For end-to-end systems, report detection F-score and recognition CER/WER (both based on Levenshtein edit distance; see CTC); for layout-heavy tasks, track IoU/tightness and character-level normalized edit distance as in ICDAR RRC evaluation kits.
  • Export rich outputs. Prefer hOCR /ALTO (or both) so you keep coordinates and reading order—vital for search hit highlighting, table/field extraction, and provenance. Tesseract’s CLI and pytesseract make this a one-liner.

Looking ahead

The strongest trend is convergence: detection, recognition, language modeling, and even task-specific decoding are merging into unified Transformer stacks. Pretraining on large synthetic corpora remains a force multiplier. OCR-free models will compete aggressively wherever the target is structured outputs rather than verbatim transcripts. Expect hybrid deployments too: a lightweight detector plus a TrOCR-style recognizer for long-form text, and a Donut-style model for forms and receipts.

Further reading & tools

Tesseract (GitHub) · Tesseract docs · hOCR spec · ALTO background · EAST detector · OpenCV text detection · TrOCR · Donut · COCO-Text · SynthText · Kraken · Calamari OCR · ICDAR RRC · pytesseract · IAM handwriting · OCR file-format tools · EasyOCR

Frequently Asked Questions

What is OCR?

Optical Character Recognition (OCR) is a technology used to convert different types of documents, such as scanned paper documents, PDF files or images captured by a digital camera, into editable and searchable data.

How does OCR work?

OCR works by scanning an input image or document, segmenting the image into individual characters, and comparing each character with a database of character shapes using pattern recognition or feature recognition.

What are some practical applications of OCR?

OCR is used in a variety of sectors and applications, including digitizing printed documents, enabling text-to-speech services, automating data entry processes, and assisting visually impaired users to better interact with text.

Is OCR always 100% accurate?

While great advancements have been made in OCR technology, it isn't infallible. Accuracy can vary depending upon the quality of the original document and the specifics of the OCR software being used.

Can OCR recognize handwriting?

Although OCR is primarily designed for printed text, some advanced OCR systems are also able to recognize clear, consistent handwriting. However, typically handwriting recognition is less accurate because of the wide variation in individual writing styles.

Can OCR handle multiple languages?

Yes, many OCR software systems can recognize multiple languages. However, it's important to ensure that the specific language is supported by the software you're using.

What's the difference between OCR and ICR?

OCR stands for Optical Character Recognition and is used for recognizing printed text, while ICR, or Intelligent Character Recognition, is more advanced and is used for recognizing hand-written text.

Does OCR work with any font and text size?

OCR works best with clear, easy-to-read fonts and standard text sizes. While it can work with various fonts and sizes, accuracy tends to decrease when dealing with unusual fonts or very small text sizes.

What are the limitations of OCR technology?

OCR can struggle with low-resolution documents, complex fonts, poorly printed texts, handwriting, and documents with backgrounds that interfere with the text. Also, while it can work with many languages, it may not cover every language perfectly.

Can OCR scan colored text or colored backgrounds?

Yes, OCR can scan colored text and backgrounds, although it's generally more effective with high-contrast color combinations, such as black text on a white background. The accuracy might decrease when text and background colors lack sufficient contrast.

What is the JPEG format?

Joint Photographic Experts Group JFIF format

JPEG, which stands for Joint Photographic Experts Group, is a commonly used method of lossy compression for digital images, particularly for those images produced by digital photography. The degree of compression can be adjusted, allowing a selectable tradeoff between storage size and image quality. JPEG typically achieves 10:1 compression with little perceptible loss in image quality.

The JPEG compression algorithm is at the core of the JPEG standard. The process begins with a digital image being converted from its typical RGB color space into a different color space known as YCbCr. The YCbCr color space separates the image into luminance (Y), which represents the brightness levels, and chrominance (Cb and Cr), which represent the color information. This separation is beneficial because the human eye is more sensitive to variations in brightness than color, allowing the compression to take advantage of this by compressing color information more than luminance.

Once the image is in the YCbCr color space, the next step in the JPEG compression process is to downsample the chrominance channels. Downsampling reduces the resolution of the chrominance information, which typically doesn't affect the perceived quality of the image significantly, due to the human eye's lower sensitivity to color detail. This step is optional and can be adjusted depending on the desired balance between image quality and file size.

After downsampling, the image is divided into blocks, typically 8x8 pixels in size. Each block is then processed separately. The first step in processing each block is to apply the Discrete Cosine Transform (DCT). The DCT is a mathematical operation that transforms the spatial domain data (the pixel values) into the frequency domain. The result is a matrix of frequency coefficients that represent the image block's data in terms of its spatial frequency components.

The frequency coefficients resulting from the DCT are then quantized. Quantization is the process of mapping a large set of input values to a smaller set – in the case of JPEG, this means reducing the precision of the frequency coefficients. This is where the lossy part of the compression occurs, as some image information is discarded. The quantization step is controlled by a quantization table, which determines how much compression is applied to each frequency component. The quantization tables can be adjusted to favor higher image quality (less compression) or smaller file size (more compression).

After quantization, the coefficients are arranged in a zigzag order, starting from the top-left corner and following a pattern that prioritizes lower frequency components over higher frequency ones. This is because lower frequency components (which represent the more uniform parts of the image) are more important for the overall appearance than higher frequency components (which represent the finer details and edges).

The next step in the JPEG compression process is entropy coding, which is a method of lossless compression. The most common form of entropy coding used in JPEG is Huffman coding, although arithmetic coding is also an option. Huffman coding works by assigning shorter codes to more frequent occurrences and longer codes to less frequent occurrences. Since the zigzag ordering tends to group similar frequency coefficients together, it increases the efficiency of the Huffman coding.

Once the entropy coding is complete, the compressed data is stored in a file format that conforms to the JPEG standard. This file format includes a header that contains information about the image, such as its dimensions and the quantization tables used, followed by the Huffman-coded image data. The file format also supports the inclusion of metadata, such as EXIF data, which can contain information about the camera settings used to take the photograph, the date and time it was taken, and other relevant details.

When a JPEG image is opened, the decompression process essentially reverses the compression steps. The Huffman-coded data is decoded, the quantized frequency coefficients are de-quantized using the same quantization tables that were used during compression, and the inverse Discrete Cosine Transform (IDCT) is applied to each block to convert the frequency domain data back into spatial domain pixel values.

The de-quantization and IDCT processes introduce some errors due to the lossy nature of the compression, which is why JPEG is not ideal for images that will undergo multiple edits and re-saves. Each time a JPEG image is saved, it goes through the compression process again, and additional image information is lost. This can lead to a noticeable degradation in image quality over time, a phenomenon known as 'generation loss'.

Despite the lossy nature of JPEG compression, it remains a popular image format due to its flexibility and efficiency. JPEG images can be very small in file size, which makes them ideal for use on the web, where bandwidth and loading times are important considerations. Additionally, the JPEG standard includes a progressive mode, which allows an image to be encoded in such a way that it can be decoded in multiple passes, each pass improving the image's resolution. This is particularly useful for web images, as it allows a low-quality version of the image to be displayed quickly, with the quality improving as more data is downloaded.

JPEG also has some limitations and is not always the best choice for all types of images. For example, it is not well-suited for images with sharp edges or high contrast text, as the compression can create noticeable artifacts around these areas. Additionally, JPEG does not support transparency, which is a feature provided by other formats like PNG and GIF.

To address some of the limitations of the original JPEG standard, new formats have been developed, such as JPEG 2000 and JPEG XR. These formats offer improved compression efficiency, support for higher bit depths, and additional features like transparency and lossless compression. However, they have not yet achieved the same level of widespread adoption as the original JPEG format.

In conclusion, the JPEG image format is a complex balance of mathematics, human visual psychology, and computer science. Its widespread use is a testament to its effectiveness in reducing file sizes while maintaining a level of image quality that is acceptable for most applications. Understanding the technical aspects of JPEG can help users make informed decisions about when to use this format and how to optimize their images for the balance of quality and file size that best suits their needs.

Supported formats

AAI.aai

AAI Dune image

AI.ai

Adobe Illustrator CS2

AVIF.avif

AV1 Image File Format

BAYER.bayer

Raw Bayer Image

BMP.bmp

Microsoft Windows bitmap image

CIN.cin

Cineon Image File

CLIP.clip

Image Clip Mask

CMYK.cmyk

Raw cyan, magenta, yellow, and black samples

CUR.cur

Microsoft icon

DCX.dcx

ZSoft IBM PC multi-page Paintbrush

DDS.dds

Microsoft DirectDraw Surface

DPX.dpx

SMTPE 268M-2003 (DPX 2.0) image

DXT1.dxt1

Microsoft DirectDraw Surface

EPDF.epdf

Encapsulated Portable Document Format

EPI.epi

Adobe Encapsulated PostScript Interchange format

EPS.eps

Adobe Encapsulated PostScript

EPSF.epsf

Adobe Encapsulated PostScript

EPSI.epsi

Adobe Encapsulated PostScript Interchange format

EPT.ept

Encapsulated PostScript with TIFF preview

EPT2.ept2

Encapsulated PostScript Level II with TIFF preview

EXR.exr

High dynamic-range (HDR) image

FF.ff

Farbfeld

FITS.fits

Flexible Image Transport System

GIF.gif

CompuServe graphics interchange format

HDR.hdr

High Dynamic Range image

HEIC.heic

High Efficiency Image Container

HRZ.hrz

Slow Scan TeleVision

ICO.ico

Microsoft icon

ICON.icon

Microsoft icon

J2C.j2c

JPEG-2000 codestream

J2K.j2k

JPEG-2000 codestream

JNG.jng

JPEG Network Graphics

JP2.jp2

JPEG-2000 File Format Syntax

JPE.jpe

Joint Photographic Experts Group JFIF format

JPEG.jpeg

Joint Photographic Experts Group JFIF format

JPG.jpg

Joint Photographic Experts Group JFIF format

JPM.jpm

JPEG-2000 File Format Syntax

JPS.jps

Joint Photographic Experts Group JPS format

JPT.jpt

JPEG-2000 File Format Syntax

JXL.jxl

JPEG XL image

MAP.map

Multi-resolution Seamless Image Database (MrSID)

MAT.mat

MATLAB level 5 image format

PAL.pal

Palm pixmap

PALM.palm

Palm pixmap

PAM.pam

Common 2-dimensional bitmap format

PBM.pbm

Portable bitmap format (black and white)

PCD.pcd

Photo CD

PCT.pct

Apple Macintosh QuickDraw/PICT

PCX.pcx

ZSoft IBM PC Paintbrush

PDB.pdb

Palm Database ImageViewer Format

PDF.pdf

Portable Document Format

PDFA.pdfa

Portable Document Archive Format

PFM.pfm

Portable float format

PGM.pgm

Portable graymap format (gray scale)

PGX.pgx

JPEG 2000 uncompressed format

PICT.pict

Apple Macintosh QuickDraw/PICT

PJPEG.pjpeg

Joint Photographic Experts Group JFIF format

PNG.png

Portable Network Graphics

PNG00.png00

PNG inheriting bit-depth, color-type from original image

PNG24.png24

Opaque or binary transparent 24-bit RGB (zlib 1.2.11)

PNG32.png32

Opaque or binary transparent 32-bit RGBA

PNG48.png48

Opaque or binary transparent 48-bit RGB

PNG64.png64

Opaque or binary transparent 64-bit RGBA

PNG8.png8

Opaque or binary transparent 8-bit indexed

PNM.pnm

Portable anymap

PPM.ppm

Portable pixmap format (color)

PS.ps

Adobe PostScript file

PSB.psb

Adobe Large Document Format

PSD.psd

Adobe Photoshop bitmap

RGB.rgb

Raw red, green, and blue samples

RGBA.rgba

Raw red, green, blue, and alpha samples

RGBO.rgbo

Raw red, green, blue, and opacity samples

SIX.six

DEC SIXEL Graphics Format

SUN.sun

Sun Rasterfile

SVG.svg

Scalable Vector Graphics

TIFF.tiff

Tagged Image File Format

VDA.vda

Truevision Targa image

VIPS.vips

VIPS image

WBMP.wbmp

Wireless Bitmap (level 0) image

WEBP.webp

WebP Image Format

YUV.yuv

CCIR 601 4:1:1 or 4:2:2

Frequently asked questions

How does this work?

This converter runs entirely in your browser. When you select a file, it is read into memory and converted to the selected format. You can then download the converted file.

How long does it take to convert a file?

Conversions start instantly, and most files are converted in under a second. Larger files may take longer.

What happens to my files?

Your files are never uploaded to our servers. They are converted in your browser, and the converted file is then downloaded. We never see your files.

What file types can I convert?

We support converting between all image formats, including JPEG, PNG, GIF, WebP, SVG, BMP, TIFF, and more.

How much does this cost?

This converter is completely free, and will always be free. Because it runs in your browser, we don't have to pay for servers, so we don't need to charge you.

Can I convert multiple files at once?

Yes! You can convert as many files as you want at once. Just select multiple files when you add them.