JPG to Text Converter

Taking a photo that reads well

OCR accuracy depends far more on the photo than on the software. Hold the phone parallel to the page so lines stay straight, and fill the frame with the text you need. Daylight from a window, or a lamp from the side, avoids the hard shadow your hand and phone cast when lit from above. Tap to focus on the text and keep still for a moment; motion blur is the most common cause of garbled words.

A flat page beats a curved one: press a book open with your free hand or photograph one page at a time. Glossy paper can reflect a bright spot that wipes out a few words, so tilt slightly until the glare disappears.

Receipts, menus and labels

Receipts are a good fit for OCR because they use plain fonts in a single column. Keep “Keep line breaks” on so each item and price stays on its own line, which makes it easy to paste into a spreadsheet or an expense report. Thermal receipts fade over time, so photograph them early; faint gray print loses letters.

Menus, product labels and signs often mix fonts, colors and decorative lettering. Plain text in them is usually read well, while stylized headings may be skipped or misread.

JPG quality and resolution

JPG compression can blur fine edges, but photos straight from a phone camera are saved at high quality and read fine. Images that have been forwarded through chat apps are often shrunk and heavily compressed, so send yourself the original if you can. If a photo is much larger than the text you need, cropping it first with Crop Image helps the engine focus and is quicker.

Small images are enlarged automatically before reading. Enlarging cannot invent detail that isn’t there, though, so a tiny, blurry thumbnail will still produce errors.

Turning the text into something useful

Once the words are out of the photo, they behave like any other text. Paste a recipe into your notes app, search a long letter for a date, translate a menu with your usual translation app, or drop an address straight into a map. For paperwork, keeping line breaks preserves the shape of forms and receipts; turning them off gives clean sentences for letters and articles. Each photo also gets its own .txt file, named after the original picture, so a batch of scanned pages stays in order when you download them as a ZIP.

Private by default

Photos of documents often hold names, addresses and account numbers. The text recognition runs entirely in your browser, so the photo and the extracted text stay on your device. Only the recognition engine and the language file are downloaded, once, from our own site. For more on image formats, read JPG vs PNG vs WebP vs AVIF.