Decode Anything

How the ranking works

Running a hundred decoders over one string is easy; the hard part is that ninety-nine of the answers are wrong, and reading them all is slower than trying the ciphers by hand would have been. So each reading is scored twice: once for how much it reads like English, and once for whether it contains something shaped like a coordinate.

That second score is the one that matters here. A mystery cache answer is far more often N 51 24.570 W 000 12.345 than it is a sentence, and a bare coordinate scores terribly as English (it's mostly digits). General-purpose cipher tools bury it near the bottom for exactly that reason. A reading only has to look like one of the two to come to the top here.

What it tries

All 25 Caesar shifts, Atbash, ROT47, Baconian in both alphabets and both polarities, A1Z26, character codes in decimal and octal, binary in 7 and 8 bits, Polybius, hexadecimal, base32, base58, base64, Ascii85, Morse, phone multi‑tap, the NATO alphabet, keyboard shifts, reversal and URL escaping.

That's a curated list rather than an exhaustive one. If your string needs a keyword (Vigenère, Playfair, a keyed substitution), nothing here will find it, because there's nothing to brute-force without the key. Try the letter frequencies first: an index of coincidence near 0.067 says the alphabet has merely been swapped about, and near 0.040 says a keyword is involved.

If nothing looks right

Decodings often stack. Click Use as input on a promising but still unreadable reading to run everything over it again: base64 wrapped round a ROT13 is a common pairing. It's also worth pasting the whole cache description into the coordinate sniffer, in case the answer was never encoded at all and is simply sitting there in plain sight.