The built-in codebook includes homophones; an imported book may not. This controls whether encoding picks randomly from the pool or always uses the first code
The built-in codebook includes null codes; an imported book may not. This controls whether encoding inserts them at random positions
Additive is the strongest option — with a unique random pad per message it is theoretically unbreakable. Reusing a pad across messages destroys that guarantee.
0 chars
⚠ This pad has been used once — generate a new one for the next message.Save the pad before sending the message! The pad — not the passphrase — is what decodes additive ciphertext. Leave this empty to fall back to a passphrase-derived stream (no longer a true OTP).
Plaintext inputprintable ASCII + newlines
Encrypted output
In0
chars
Out0
chars
–Phrases
–Words
–Trigrams
–Digrams
–Letters
–Nulls
–Ratio
Live encodingclick Run to watch each token become a code group
Encoding mapclick any token or code group to highlight its match —
click in the input or output textareas above to find the corresponding span here
Plaintext (input)
Encode or decode a message to see the breakdown.
Code groups (output)
phrasewordtrigramdigramletterspecialnull
Encoding trace
run an encode to see results
Phrases
Words
Trigrams
Digrams
Letters
Special
A→Z
Import Codebook
Load a JSON or TSV codebook. The imported book will replace the current one. JSON format:{"seed":"…","groupLen":5,"entries":{"code":"token",…}} TSV format: one code TAB token pair per line (first line may be a #seed:… comment).
Or paste below. A paste over 1 MB is checked straight away instead of being shown in the box (a box that full would freeze the tab); the file picker above is still the easier way for codebooks generated with high homophone multipliers or the full word list. Any book the CODEX Maker makes will load (the import takes up to 85 MB and 3,500,000 codes); the biggest can take several seconds.