Steganography detector

Checks an image for traces of steganography — data hidden in the least-significant bits (LSB) of the pixels. LSB stego is the most common class: a payload (text, file, key) replaces one bit per RGB channel in every pixel; the image looks unchanged but the statistical distribution of the bottom bits stops being random. This tool runs the Westfeld–Pfitzmann chi-square test (1999) — the same method used by academic detectors like StegExpose, StegDetect, Stegano. For each R/G/B channel it builds a histogram, groups adjacent pairs (2k, 2k+1), compares observed frequencies to the expected midpoint, and computes a chi-square p-value. A high p-value (>0.5) is suspicious: the pairs are “too even”. On top of the numbers, the tool also renders an LSB map — a black-and-white image where each pixel is the least-significant bit of the original. On a clean photo this looks like uniform noise; on a stego-loaded image you can often see the payload’s structure with the naked eye — text, geometry, regular patterns. Limitations are stated up front: JPEG normally destroys LSB on re-save (DCT quantisation wipes the bottom bits), so the test is most reliable on PNG and BMP, and only triggers on JPEG if the file was stego-processed exactly once and never re-saved. A positive result is a reason to verify with proper forensic tools; a negative one does not rule out more sophisticated stego (DCT-stego in JPEG, F5, OutGuess). No vendors, plain JavaScript. Processing happens in the browser, the file never leaves your device.