Same name. Same size.
Same file?

Each is a different question. A useful duplicate finder keeps those distinctions visible so you can decide with the right evidence.

A filename is a label

Two files named invoice.pdf may belong to different months. A renamed copy can contain exactly the same bytes under a different filename. Neither matching names nor different names settles whether the contents match. Folders provide context; they do not change a file's bytes.

Size is a fast first check

Byte-identical files must have the same length. Grouping non-empty files by size avoids reading every file for a hash. But the converse does not hold: two files can both be 1,024 bytes and contain entirely different data. Size-only groups belong in a candidate list until content evidence is available.

A hash is a content fingerprint

SHA-256 derives a 256-bit value from a file's bytes. Changing the bytes normally changes the fingerprint. Duplicate File Locator hashes eligible same-size candidates and groups those with matching size and SHA-256. Filenames and modification dates are not part of this hash.

A matching cryptographic hash is extremely strong evidence, rather than a mathematical guarantee that no collision is possible. This app does not perform a second byte-by-byte comparison of every matched pair. Its “Exact copies” label means same-size and matching-SHA-256 evidence within the completed scan.

Why verification has a limit

The browser's Web Crypto digest reads an entire input buffer instead of accepting a stream. This app limits duplicate hashing to files of 128 MiB or less and processes them sequentially. That limits the size of each hashing operation, but it cannot guarantee that every device has enough memory.

Files over the limit, failed reads, and unavailable hashing stay in Possible copies. They are not included in verified extra-copy byte estimates. A scan with no verified groups can still contain unverified candidates. Empty files are left out of duplicate detection because they contain no payload, although you can find them in Files.

What a hash cannot tell you

A hash does not establish who created a document, whether an image is authentic, or which location you should keep. It also does not compare the meaning of documents. Two ZIP archives containing equivalent extracted files may differ in container metadata or compression, so this tool compares the archives as whole byte sequences.

A scan describes bytes readable at the time of processing. If another application changes the source files afterwards, reconnect the folder and verify again before taking action elsewhere.

A useful way to read results

• Exact copies: content evidence matched; inspect where and why the copies exist.
• Possible copies: some verification is missing; keep uncertainty in your decision.
• Name conflicts: names match; read the files to understand whether they are different.

For the implementation details behind browser hashing, see MDN's Web Crypto digest documentation. For the app's workflow, read the manual.