Binary Code Translator from Image
Extract printed 0s and 1s from a screenshot or photo, check the recognition, and decode the resulting text bytes. OCR runs in your browser. This reads visible binary text; it does not convert image pixels or file bytes into binary.
On this page
Try a two-byte practice image
The sample contains 01001000 01101001, which decodes to Hi in ASCII or UTF-8. It is a clean synthetic teaching fixture, not a benchmark of real-world OCR accuracy. Download it and use the same image in the uploader above.
- Download the PNG, then choose it in the image uploader.
- Keep the whole line visible and select Recognize bits.
- Compare the editable result with 01001000 01101001. Correct any missing or mistaken digit.
- Select Translate checked bits. The expected text is Hi.
One wrong digit can still look valid
If the second byte is read as 01101000 instead of 01101001, the result is Hh rather than Hi. Both are valid ASCII. This is an illustrative error, not a claimed OCR output from the sample.
Confidence highlights areas to inspect; it cannot certify a message. Compare the image with the editable bits and preserve the intended order of lines before translating.
| Stage | Bits | Text |
|---|---|---|
| Known source | 01001000 01101001 | Hi |
| Illustrative one-bit mistake | 01001000 01101000 | Hh |
| After checking the last bit | 01001000 01101001 | Hi |
Make the digits easier to recognize
Crop away unrelated text. Rotate sideways text upright, keep a small clear margin around the digits, and try contrast or threshold controls for an uneven background. Invert light-on-dark text if recognition is poor.
Only recognized 0s, 1s, and whitespace are kept. That filter cannot recover an omitted digit. Handwriting, blur, compression, overlapping text, and multi-column layouts need particular care.
Privacy and practical limits
PNG, JPG, and WebP files are accepted up to 10 MB. Images above 40 megapixels are rejected; accepted images are scaled to at most 2,400 pixels per edge. Crop very large screenshots first so small digits remain readable.
First use downloads this site's OCR worker, WebAssembly engine, and English model. Recognition takes device memory and CPU. A camera choice depends on your mobile browser; the site does not request a live camera stream.
The image and detected text are not sent to a conversion server. You can cancel recognition and edit the bits before translating. This OCR workbench decodes eight-bit UTF-8; take a seven-bit or five-bit source to the matching converter.
Frequently asked questions
How do I translate binary from a screenshot or photo?
Choose, drop, or paste an image containing visible 0s and 1s, then select Recognize bits. Compare the editable bits with the image and correct mistakes before selecting Translate checked bits. The built-in decoder reads eight-bit UTF-8. Start with the downloadable Hi image if you want a known result to check the workflow.
PermalinkWhich image formats and sizes are supported?
The uploader accepts PNG, JPG, and WebP, with a file-size limit of 10 MB (10 × 1,024 × 1,024 bytes). Images above 40 megapixels are rejected, and accepted images are scaled to at most 2,400 pixels per edge. Crop oversized screenshots before choosing the file so small digits survive resizing. PDF and GIF files are not accepted directly.
PermalinkIs my image uploaded to an OCR service?
No. The browser downloads the OCR worker, engine, and language model from this site, then processes your selected image on your device. The image and recognized text are not sent to a conversion server. This does not mean the page makes no network requests: the recognition software must be available before OCR can run.
PermalinkHow can I improve recognition of blurry or low-contrast digits?
Use a sharper source image when possible, crop out unrelated text, and rotate the digits upright. Then try contrast, threshold, or inversion controls while checking that 0s and 1s remain distinct. Keep some clear margin around the text. Filters cannot reconstruct a missing digit, and handwriting or multi-column layouts may still need manual transcription.
PermalinkCan I trust a high confidence score or an unmarked region?
A confidence score is a review aid, not a correctness guarantee. The preview outlines recognized regions below 80% confidence, but unmarked digits can also be wrong. Changing the final bit of 01101001 to 0 produces h instead of i, and both are valid text. Compare all digits and their reading order with the original.
PermalinkWhy must I translate again after correcting the recognized bits?
Editing the detected bits clears the previously approved result so it cannot be mistaken for a translation of your new input. After checking the edit against the image, select Translate checked bits again. If an incomplete-byte error remains, count the source digits and check the encoding; the tool does not invent missing bits.
PermalinkCan the image tool decode seven-bit ASCII, five-bit codes, or numbers?
OCR can help extract the visible digits, but this workbench's Translate checked bits action uses eight-bit UTF-8 only. For a known seven-bit ASCII source, copy the checked bits to the text decoder and select seven-bit ASCII. For A1Z26, ITA2, or an unsigned number, use the corresponding tool instead of adding padding to force UTF-8.
PermalinkCan I use my phone camera or paste an image?
A mobile browser may offer a camera option through the file chooser; availability depends on the browser and device. This page does not request a live camera stream. On a compatible desktop browser, drop an image into the panel or focus that panel and paste an image from the clipboard. Choosing a saved supported file is the fallback.
PermalinkWhy is the first recognition slow, and can I cancel it?
First recognition loads the OCR software and English model, then uses your device's CPU and memory to analyze the image. Download speed and image complexity affect the wait. Use Cancel recognition to stop an active run; you can also type the bits manually. Offline operation is not guaranteed because the required assets may not be cached.
PermalinkCan I decode QR codes, hidden messages, or a photo's pixels here?
No. This tool recognizes printed 0s and 1s that you can see in an image. It does not read QR symbols, steganography, image-file bytes, or pixel color encodings. If you want to create a picture of a binary message from ordinary text, use the binary graphic generator rather than the image reader.
Permalink↑ Back to the converter · Page contents
Sources & verification
Reference material used for this explanation. The examples are reproducible with the tools on this site; see how they are checked.
Found a mismatch? Send a correction with a source.