QR Code Reader

Features

• Read a code from a screenshot or photo, including inverted codes (light on dark)
• Recognises the payload type: URL, WiFi, vCard, calendar, email, phone, SMS, coordinates
• Splits WiFi settings into network, security and password — escaping handled correctly
• Reports the QR version and module count, and the encoding used (numeric, alphanumeric, byte, kanji)
• Two passes: a fast downscaled scan, then full resolution if that fails
• Several images at once, one result card each — an unreadable one still gets a card naming the file and the reason, so nothing silently disappears
• Copy the result, or download it as text
• Nothing is uploaded and no server is involved

How to use

1. Drop an image containing a code in (a screenshot or a photo)
2. The result appears at once, with the breakdown underneath
3. Use Copy to take the content, or download it as text
4. If nothing is found, crop to just the code and try again

FAQ

It does not find the code. What now?

In this order: crop away the background (a busy one interferes with locating the code), use a sharper image, scale it up (a code in a screenshot is often only a couple of hundred pixels). Also note that glare, a steep viewing angle and watermarks all defeat recognition — a human eye fills in the gaps, an algorithm cannot. The tool already scans a downscaled copy first and retries at full resolution, so you do not have to tune anything yourself.

Is the WiFi password shown correctly?

Yes, and that is the most useful part of this tool. A WiFi code reads `WIFI:T:WPA;S:name;P:password;;`, which looks simple — but a semicolon, colon or backslash inside the name or password is escaped as `\;` and `\:`, so splitting on semicolons truncates the password. The user then has a wrong password and assumes they mistyped it. This parses the escaping properly, so special characters come through intact.

What if the image holds several codes?

Only one is read (the algorithm returns the first code it can locate). To read them all, drop several images in at once — each gets its own card — or crop them apart first. That matches what you actually want better than the tool guessing which ones matter. Note too that if an image has both a QR code and a barcode, this tool only handles the QR code.

What if I drop in many images and some fail?

A failed image still gets a card, showing the file name and “no QR code found”, so you can see exactly which ones were the problem while the rest come through normally. That is deliberate: if the failures simply vanished you would see a screen of results and never count how many there should have been.

Is the image uploaded?

No. Decoding runs a local JS library over the image pixels — binarising and locating — entirely in browser memory with no network requests, and it keeps working offline.