
Magic decode
Try base64, base32, hex, URL-encoding, ROT13/47 and single-byte XOR against a string and rank the results by how much they read like real text.
Magic decode throws your text at seven guesses at once: base64, base32, hex, URL-decoding, ROT13, ROT47, plus a brute-forced single-byte XOR. Each result is scored on how much it reads like real text, blending printable-character ratio, common-word hits and letter frequency into one number. XOR brute force checks only the first 4096 bytes. A high score is a lead, not proof.
Questions
What decodings does it try?
Base64, base32, hex, URL-decoding, ROT13, ROT47 and a brute-forced single-byte XOR, all against the same input text.
How is each candidate scored?
A weighted mix: 40% printable-character ratio, 30% common-English-word hits, and 30% letter-frequency fit against English.
How much of the input does XOR brute force check?
Only the first 4096 bytes, and it keeps the single best-scoring key out of all 255 tried, not every result.
Does a high score prove the encoding was correct?
No, it is a ranked lead, not proof; the text score can be fooled by coincidentally text-like garbage.
Does it try multi-byte XOR keys?
No, only single-byte keys are brute forced.