Paste any unknown string, ciphertext or payload. Every decoder we have runs at once — encodings, compression, classical ciphers and brute-force families — and the results are ranked by how much they look like real language. Optional key, number and alphabet hints unlock the keyed ciphers.
The text's index of coincidence is high, meaning one letter likely maps to one letter, but the alphabet is scrambled.
Everything at once. One paste runs every decoder in the library: Base64 / Base32 / Base45 / Base58 / Base62 / Ascii85 / Z85, hex, binary, octal, decimal, URL and double-URL encoding, HTML entities, Quoted-Printable, Unicode escapes, Punycode, JWT, Gzip and Deflate, Morse (international and Russian), Braille, tap code, NATO, Baconian, A1Z26, T9 multi-tap, leetspeak and reversal.
Brute-force families. Caesar runs all 25 Latin (32 Cyrillic) shifts, Affine all 311 valid a·x+b pairs, rail fence 2–10 rails, keyboard shift ±2 keys, and single-byte XOR all 255 keys when the input decodes to bytes. Affine and XOR keep only their best-scoring hits — the remaining few hundred are rotations of the same noise and only make the list harder to read.
Automatic key recovery. For Vigenère, Beaufort, Variant Beaufort, Gronsfeld and Porta the keyword is worked out for you, no hint needed. The letters are split into columns to find the key length by index of coincidence, each column is matched against the language's letter frequencies, and the result is then hill-climbed on letter-pair and word statistics to fix the columns the frequencies got wrong. Give it roughly 40 letters or more — below that there is not enough signal, and Porta needs the most of all.
Key hints. If you already know the keyword, type it and Vigenère, Beaufort, Variant Beaufort, Autokey, Porta, Gronsfeld, Playfair, Polybius, Bifid (periods 3–7), ADFGX, ADFGVX and columnar transposition are all tried with it. Multiple keywords can be given, separated by spaces.
Ranking. Each candidate is scored 0–100 on dictionary words, letter-pair frequencies, JSON/URL/HTML structure and the share of unprintable characters. The three most plaintext-like readings are lifted into Top matches at the top of the results. It is a heuristic, not proof — always eyeball the output.
Grouping. Results are grouped by decoder, not by variant: Caesar's 25 shifts, Affine's 312 a·b pairs and both Baconian readings each collapse into one card showing the best-scoring variant, with the rest one click away. Decoders that landed on the same text — ignoring case, spacing and punctuation — are merged too, and the clearest rendering keeps the card while the others are listed under “same text from”.
Cascade. When an output is itself encoded, decoding continues automatically up to five layers deep (for example Base64 → Gzip → JSON), and the chain is shown at the top.
Coordinates. Any decoded result containing coordinates is flagged with a map link — useful for geocaching and puzzle caches. Alongside DD, DDM (N 51° 28.123) and DMS, the projected grids are converted too: UTM (30U 699319 5710158), MGRS (31U DQ 48251 11932) and the British National Grid (TQ 38863 77161), the last of which is shifted onto WGS84 so it lands where your GPS expects rather than 100 m away.
Everything runs locally in your browser. Nothing you paste is uploaded, logged or stored.
