BOM Detector

Some text files begin with a byte order mark, three to four bytes declaring the encoding before the first real character. It is effectively mandatory for UTF-16 and UTF-32, where it indicates byte order, and optional in UTF-8, where it merely hints. Since those bytes are invisible in any editor, a file with an encoding problem gives no clue at all on screen.

Pick the file and the page reads its beginning and identifies the mark: UTF-8, UTF-16 in either order, or UTF-32 in either order. When there is no mark, it says so — which does not mean the file is UTF-8, only that it does not declare itself. The file is read in your own browser and is not sent anywhere.

The order of the checks matters more than it seems. The UTF-32 little-endian mark starts with the same two bytes as the UTF-16 little-endian mark, so testing UTF-16 first would misclassify every UTF-32 file. That is why the four-byte patterns are checked before the two-byte ones — a detail easy to get wrong in a hand-rolled implementation.