Select several files to hash them in the browser and group the ones that are byte-for-byte identical, with the space the extra copies take up.
The finder reads each selected file and computes its SHA-256 hash with the browser's cryptography API, one file at a time so progress can be shown. Files sharing a hash are byte-for-byte identical — the comparison uses content, not names, so renamed copies are still caught, and two different files matching by hash is not a realistic possibility with SHA-256. Each duplicate group lists its files and size, and the reclaimable total counts every copy beyond the first in each group. Browsers expose only file names to a page, never full paths, so groups are identified by name and size.
Are the files uploaded for hashing? No. Files are hashed in the browser with the Web Crypto API and never uploaded. The page sees file names and contents locally only, and nothing is stored.
Can two different files have the same SHA-256 hash? In principle any hash function can collide, but no two different files with the same SHA-256 are known. For duplicate finding, equal SHA-256 hashes are treated as identical content.
Why are files identified by name rather than folder? Browsers deliberately do not reveal a file's path to a web page, for privacy. Names and sizes are shown so copies can be told apart where names differ; same-name copies in one group have the same size and content.