QR Code Generator

A QR code generator turns text, URLs, contact details, or other short payloads into a square barcode that phones and scanners can read quickly. QR codes are useful on posters, labels, menus, product packaging, table cards, internal signs, event check-ins, and any place where typing a link would be inconvenient. A good QR code is not only generated, but also scannable: contrast, size, quiet zone, print quality, viewing distance, and the amount of encoded data all matter. Error correction helps a code survive small damage or visual decoration, but higher correction can also make dense content harder to scan. Before publishing, always scan the exported PNG or SVG with a phone under real lighting conditions, especially after changing colors or using a long URL.

QR Code is standardized as ISO/IEC 18004:2015 and defines 40 size versions: Version 1 is a 21 × 21 module matrix and each subsequent version adds 4 modules per side, so Version 40 is 177 × 177. A 'module' is one black-or-white square in the grid. Three fixed 7 × 7 finder patterns sit at the top-left, top-right, and bottom-left corners and let a scanner lock orientation regardless of rotation; from Version 2 onward, smaller 5 × 5 alignment patterns are added on a regular grid to correct perspective distortion, and a one-module-wide timing pattern runs between the finders to establish the module pitch.<br /><br />Payload bytes are wrapped with Reed-Solomon error correction over GF(2^8), which is what lets a partially obscured code still decode. The four levels are L = recover up to 7 percent of the codewords, M = 15 percent, Q = 25 percent, and H = 30 percent. Choosing H lets you overlay a centered logo of roughly 20 percent of the code area without losing readability, at the cost of fewer usable data codewords for the same version. Data itself is encoded under one of four modes — Numeric (10 bits per 3 digits), Alphanumeric (11 bits per 2 chars from a 45-symbol set), Byte (8 bits, typically UTF-8 for non-ASCII), and Kanji (13 bits per Shift_JIS character) — and the encoder picks the densest mode that covers every input character.<br /><br />After encoding, the bit stream is XORed with one of eight mask patterns to avoid large monochrome regions and the false detection of finder-like sequences; the encoder evaluates all eight against a penalty score (defined in section 7.8.3 of the standard) and writes the winning mask ID into the format-information stripes next to the finders. Format and version metadata are themselves protected by BCH(15,5) and BCH(18,6) codes. In the browser, the qrcode npm library generates the matrix and renders it onto an HTML <canvas> via CanvasRenderingContext2D.fillRect or emits SVG <rect> elements; on supported devices, the BarcodeDetector API (Chrome on Android, Safari since 16.4) decodes back from camera frames without external libraries.

• Standard: ISO/IEC 18004:2015 — 40 versions, Version N is (4N+17) × (4N+17) modules, so V1=21×21, V10=57×57, V40=177×177
• Finder and alignment patterns: three fixed 7×7 finders at TL/TR/BL for rotation lock; 5×5 alignment patterns from V2+ on a regular grid for perspective correction; 1-module timing pattern between finders sets module pitch
• Error correction (Reed-Solomon over GF(2^8)): L=7%, M=15%, Q=25%, H=30% of codewords recoverable — H tolerates ~20% logo overlay at the centre
• Data modes and bit cost: Numeric 10 bits / 3 digits, Alphanumeric 11 bits / 2 chars (45-symbol set), Byte 8 bits (UTF-8 in practice), Kanji 13 bits / char (Shift_JIS)
• Capacity examples (V40-L): 7089 numeric, 4296 alphanumeric, 2953 byte, 1817 Kanji characters — capacity drops sharply at level H
• Masking: 8 mask patterns scored against a 4-rule penalty (ISO/IEC 18004 §7.8.3); the winning mask ID is written into the format-info stripes protected by BCH(15,5); version info uses BCH(18,6)
• Browser rendering and decoding: qrcode.js / qrcode (npm) emit either canvas fillRect or SVG <rect>; BarcodeDetector API (Chrome on Android, Safari 16.4+) decodes QR from camera frames without a JS library