Convert JPG to text in your browser
Most phone cameras save JPEG by default, so a jpg to text converter is often the practical path from a photo of a whiteboard, menu, or document to something you can paste elsewhere. Photo2Text accepts JPG and JPEG files, checks size on the client, optionally downscales very wide images, and runs best-effort OCR with Tesseract, without sending the picture to our servers.
JPEG-specific notes for OCR
JPEG is lossy. Aggressive compression introduces blockiness around character edges, which hurts engines more than soft blur alone. If you re-save a screenshot as a tiny JPG for messaging apps, OCR quality can collapse. Prefer the original camera file or export at high quality. For UI captures with crisp fonts, PNG to Text is often a better pipeline.
Language choices
Default is English. Switch to the Urdu converter or Arabic image to text when the photo is in those scripts. For mixed pages, use the combined language modes on the tool above.
Related workflows: PNG to text for sharp UI fonts, screenshot to editable text, and the main image to text converter. Beginners should start with how to extract text from an image; Urdu phone photos also benefit from Urdu OCR prep tips.
How JPEG compression affects text
JPEG is a lossy format, which means it saves space by throwing away image detail that the human eye barely notices. That trade works well for photographs of landscapes and faces, but it is hard on text. Letters are made of sharp edges and thin strokes, exactly the kind of detail JPEG compression softens first. At high quality settings the damage is small and OCR reads the page well. At low quality settings the compression breaks each small block of the image into visible squares, and the edges of letters turn blocky, wavy, or doubled. Small print suffers first, because a tiny letter has only a few pixels to begin with, and losing even a few of them changes the shape.
The practical lesson is to protect the original file. Every time a JPG is opened, edited, and saved again, it is compressed again, and the damage adds up. A photo that has been saved five times looks worse than the same photo saved once, even at the same quality setting. If you need to crop or rotate a document photo, do it once and keep the result, rather than passing the file through several apps. When you export from a scanner or a camera app, choose the highest quality option for pages with text, and check the result at full zoom: if the letters look clean to your eye, the OCR engine will usually read them well.
Phone camera tips for readable JPG text
Most JPG text starts life as a phone photo, so the camera technique decides the result before OCR even starts. Hold the phone parallel to the page, directly above it, so the lines stay level and the text does not lean to one side. Bright, even daylight is the best light: a window on a cloudy day beats a harsh lamp, because a single lamp casts your shadow across the page and leaves one side dark. Tap the screen to focus on the middle of the text, hold the phone steady with both hands, and let the camera settle before you shoot. A slight shake blurs exactly the thin strokes that matter most.
Glare is the enemy of document photos. A glossy page, a laminated card, or a whiteboard reflects the light source as a bright patch that hides the words underneath. Tilt the page or move the lamp until the glare slides off the text, or photograph the page in soft daylight instead. Fill the frame with the document so the text is as large as possible, and photograph one page at a time rather than a two-page spread, because the far page is smaller and the curve near the spine bends the lines. After the shot, zoom in and check the smallest print: if you can read it clearly, the OCR engine has a fair chance too.
When JPG loses text quality
JPG loses text quality in a few predictable situations, and recognising them saves a wasted OCR run. The first is re-saving: a photo that travelled through a messaging app, a social feed, or an email attachment has usually been recompressed to a smaller size, and the copy you received is softer than the original on the sender phone. The second is a screenshot saved as a JPG. Screenshots are crisp interface text with hard edges, and JPEG compression smears those edges, so a JPG screenshot reads worse than the same screen saved as a PNG. The third is low light. A photo taken in a dim room forces the camera to raise its sensitivity, which adds grain across the whole image, and that grain sits right on top of the letters.
Tiny text is the fourth case. A distant whiteboard, a small label, or a footnote photographed from a metre away gives the engine very few pixels per letter, and no quality setting can invent detail that was never captured. Move closer or zoom optically instead of digitally, because digital zoom only enlarges the same soft pixels. If a photo has already lost its detail, converting it to PNG will not bring the detail back. The honest fix is to retake the photo with better light, a closer frame, and the highest quality setting, and to keep the original file from then on.
File size vs quality: what to choose
A JPG of a document page is usually a few hundred kilobytes at a sensible quality, and that size is a good trade: large enough to keep the letters sharp, small enough to upload and process quickly. Pushing the quality to maximum doubles or triples the file size for a small gain, while dropping to a low quality setting shrinks the file but costs you the small print. For text, err on the side of quality. A file of two or three megabytes is no problem for this tool, which accepts files up to about 10MB, and the extra detail is what the engine reads.
Very wide photos are downscaled before recognition to keep the process fast, so an enormous panorama of a whiteboard does not help as much as a closer, narrower shot of the same board. If the text spans a wide page, split it into two photos, left and right, with a little overlap, and run each one separately. When the source is already a JPG, keep it as a JPG and run OCR directly; converting it to another format adds a step without adding detail. The only time to switch formats is at the source: if you are capturing a screen, save it as PNG from the start, and if you are photographing a page, keep the camera on its highest quality JPG setting.
Try our OCR tools
Same private browser OCR. Pick the page that matches your language or file type.
- Image to Text (Home)Free private OCR for English, Urdu, and Arabic images.
- Urdu OCRExtract Urdu text from photos in your browser.
- Arabic OCRConvert Arabic images to editable text privately.
- Screenshot to TextPull text from error messages, chats, and receipts.
- PNG to TextOCR for sharp PNG screenshots and scans.