Drop any file - we read its first bytes locally and tell you the real format, even if it was renamed.
Almost every file format begins with a short, fixed run of bytes called a magic number. A PNG always starts with 89 50 4E 47, a PDF with %PDF, and any ZIP-based file (that includes DOCX and APK) with PK. The name on the file is just a label you can change; the bytes at the front are what the format really is.
Drop a file here and the tool reads only its first 64 bytes, then checks them against a table of known signatures. It all happens in your browser through the FileReader API, so the file never goes anywhere. If the signature and the file name disagree, it tells you the file was most likely renamed.
.jpg that is really a .png gets caught, because the bytes give it away..txt or .csv have no signature, so they can come back as "unknown".photo.pdf that is really a JPEG.