About OCR JPG
JPEG is the format almost every camera and phone produces, and it is lossy in a way that specifically hurts OCR. Its compression works on eight-by-eight pixel blocks in the frequency domain, discarding high-frequency detail first — and the sharp black-to-white edge of a letter stroke is exactly high-frequency detail. The result is ringing and mosquito noise around characters, which blurs the boundaries the recogniser relies on to separate one glyph from the next. Practically this means a JPEG saved at high quality recognises well while the same image saved small recognises badly, and re-saving a JPEG repeatedly compounds the damage. So the advice for JPEG input is specific: work from the largest, least-compressed original you have, never from a version that has been messaged, resized and re-saved, and if you control the capture step choose PNG for screenshots and keep JPEG quality high for photographs.
How to OCR JPG
- 01
Find the original JPEG
Use the largest, least re-saved copy you have — not one forwarded through a chat app.
- 02
Add the files
Drop in one or several JPG or JPEG images.
- 03
Select the language
Choose the matching model from the seven available.
- 04
Recognise and check
Read the output against the image, paying attention to characters that compression may have blurred together.
What this tool does
- Tuned advice for lossy JPEG input, where artefacts are the main accuracy limit
- Batch several JPEGs in one run with separate results per file
- Handles both camera photos and JPEG-exported scans
- Runs locally through WebAssembly — photos never leave the device
- Seven trained languages available offline after first use
Limitations worth knowing
Every PDF tool has constraints. Stating them plainly is more useful than discovering them halfway through your work.
- Heavily compressed or repeatedly re-saved JPEGs recognise poorly, and no setting here can restore the detail that was discarded.
- Small on-screen JPEGs — a thumbnail or a chat-app preview — usually lack the pixels needed for reliable recognition.
- Handwriting in a photo is not reliably recognised.
- Only the seven trained languages are available.
How your file is handled
This tool runs inside this browser tab, but it first downloads a recognition engine and language model — static files, fetched once and then cached by your browser. Your document is never part of that request: the engine comes down to your device, and your file stays on it. You can verify this in the Network panel, where you will see the engine assets download and no upload of your document.
Nothing is stored after the fact. Closing or reloading this tab discards the file, the result and everything derived from them, because none of it ever left your machine. Read how local processing works.
Questions about OCR JPG
Why does my JPEG recognise worse than a PNG of the same page?
JPEG compression discards high-frequency detail, which is precisely what letter edges are made of. PNG is lossless, so character boundaries stay sharp and the recogniser has more to work with.
Does resizing a JPEG before OCR help?
Enlarging does not add detail and mostly wastes time. Working from a larger original does help. If the text is tiny in the frame, re-capture rather than upscale.
Should I convert the JPEG to PNG first?
It will not recover anything — the loss already happened when the JPEG was written. Converting only prevents further degradation from additional re-saves.
Is a phone photo good enough?
Often yes, if it is well lit, shot straight on, fills the frame and was saved at high quality. A dim, angled snapshot of a whole page from a distance usually is not.
Tools that pair with this one
- OCR PNGThe format that gives OCR the least to complain about.
- OCR Image to TextPoint the recogniser at a picture and get the words it contains.
- Image to Text ConverterThe alternative to retyping what you can already see on screen.
- Scan to TextFor paper that has been scanned and now needs to be text.
- JPG to PDFTurn one or many JPG files into a PDF with the original JPEG data left untouched.
- OCR PDFTurn a scanned PDF into something you can search — without uploading it.
Read more about this
- How to OCR a PDFAbout the recognition itself: what the engine is doing to your pixels, what raises the error rate, and which problems are fixable before you run it.7 min guide
- How to make a PDF searchableAbout the artefact rather than the recognition: what a text layer is, why the page looks identical afterwards, and how to verify the result is really searchable.5 min guide