Drop in any file, disk image, or raw byte dump. Scans every offset against the same 122-entry magic-byte database Unknown File Identifier uses — not just the start of the file — and extracts whatever it finds real evidence of, using each format's actual end-of-file marker, not a guessed size.
Only formats whose header signature starts at offset 0 within the embedded data are found — a handful of entries in the database that key off a non-zero offset or footer-only signature (like TGA or DMG) can't be located this way. Reliable extraction (a real end-of-file marker, not a guessed size) is only available for PNG, JPEG, GIF, PDF, and ZIP-family files today — for everything else found, the offset and format are reported, but no download is offered, since handing back the wrong byte range is worse than not offering one. Overlapping/nested matches inside an already-claimed range are skipped once a file is successfully extracted, so a JPEG's own internal bytes won't also register as a false hit. Nothing extracted here is validated by its own dedicated tool — a carved ZIP might still be damaged internally, so run it through ZIP Inspector (or the matching format tool) next. Results are capped at 500 hits, and very large files (hundreds of MB+) may be slow since every byte offset is checked.