Numbers are distributed throughout — never appended to the end
Noun and adjective slots draw from the full word list for maximum entropy
or press Space
How entropy is calculated
Word-based
N × log₂(wordlist size). Each extra word multiplies the search space. Separator choice, capitalisation mode, and l33t substitutions add bonus bits.
Character
Length × log₂(charset size). All 4 character sets enabled gives log₂(94) ≈ 6.55 bits per character. 20 chars = 131 bits.
Sentence
Sum of log₂(pool size) per slot. Noun and adjective slots use the full word list, boosting entropy far above small grammar pools alone.
Crack time assumes 100 billion guesses/second — an offline GPU cluster attack. Online attacks are orders of magnitude slower.
All three generators are available via a REST API on Cloudflare's edge. Metadata (IP, browser, style) is logged to Discord — passwords are never logged.
POST
/api/generate{
"type": "word" | "char" | "phrase",
"count": 1–10,
// Word & phrase options
"wordCount": 2–10,
"separator": "-" | "." | "_" | " " | "~" | "!" | "@" | … | null (random),
"capMode": "first" | "last" | "random" | "vowel" | "all" | "none",
"injectNum": true | false,
"leet": true | false,
// Character options
"length": 8–64,
"lower": true | false,
"upper": true | false,
"numbers": true | false,
"special": true | false
}Response
{
"passwords": [
{
"value": "C0balt-Forg3d-7Leaps-Swiftly",
"bits": 78.4,
"strength": { "label": "Good", "level": 2 },
"time": "3.1M years",
"warnings": [],
"type": "word"
}
],
"meta": { "type": "word", "count": 1, "generatedAt": "2026-04-15T10:30:00Z" }
}cURL
curl -X POST https://YOUR_DOMAIN/api/generate \
-H "Content-Type: application/json" \
-d '{"type":"word","wordCount":6,"capMode":"random","leet":true,"count":3}'