What it does

Send a string. Get back a slug. 

curl -X POST https://x402.agentutility.ai/slugify \
  -d '{"text":"Café & Bar @ 100% — Open 24/7!","max_length":40}'
{
  "input": "Café & Bar @ 100% — Open 24/7!",
  "slug": "cafe-and-bar-at-100-percent-open-24-7",
  "mode": "kebab",
  "length": 37,
  "tokens": ["cafe", "and", "bar", "at", "100", "percent", "open", "24", "7"],
  "truncated": false
}

$0.001 USDC. Cheapest endpoint in the catalog.

Why charge for slugification at all

Reasonable question.

Most slug libraries handle the common case fine: lowercase, hyphenate, strip punctuation. But here's where they diverge:

  • José Pérez → some libraries return jose-perez, some return jose-prez (dropping the accent and the letter behind it), some leave the accents in
  • Café & Bar → some libraries return cafe-bar (dropped the &), some return cafe-and-bar (replaced it)
  • Open 24/7 → some return open-247, some return open-24-7
  • 1ère édition → behavior depends on whether the library does NFKD or NFC normalization

An agent that needs deterministic, predictable slug behavior across every input — say, for cache keys that have to match across services — can't tolerate that ambiguity. So you write your own. Or you call a paid endpoint that documents the exact behavior:

  • NFKD normalization, then combining-mark strip (José → Jose → jose)
  • Symbol replacement before normalization (& → and , % → percent , @ → at , # → hash , + → plus )
  • All non-alphanumerics collapsed to the separator
  • Word-boundary-aware truncation when max_length is set

Same input, same output, every time. Deterministic.

Four modes

kebab:          lower-kebab-case
snake:          lower_snake_case
dot:            lower.dot.case
preserve_case:  Same-Separators-But-Casing-Preserved

Default is kebab because that's what URLs and SEO conventions want. snake for Python identifiers and database column names. dot for Maven artifact IDs and Java package paths. preserve_case for filename slugs where case matters.

Unicode handling

ascii_only: true (default) strips accents and converts to ASCII. ascii_only: false keeps Unicode letters and numbers through (so 森林公园 slugs to 森林公园, and Москва slugs to москва).

The cutoff matters for non-Latin scripts. ASCII-only on Chinese or Cyrillic produces an empty slug — there's no fallback transliteration baked in. We don't pretend to know whether you want pinyin or romanization. Pass ascii_only: false and handle the script in your downstream code.

Truncation

If max_length is set and the slug exceeds it, the endpoint truncates at the nearest word boundary within the last half of the limit. So text: "the quick brown fox jumps over the lazy dog", max_length: 25 returns the-quick-brown-fox-jumps, not the-quick-brown-fox-jumps-o. The word-boundary cut preserves readability of the slug as a URL.

If no good boundary exists (one long word), we cut at the literal max.

At $0.001 the unit math is unusual

A million slug calls is $1000. That's enough that any high-volume use case should still call this. But for the low-volume case, an agent that needs five slugs is paying half a cent. Below the price point where most paid APIs even bother to publish a price.

We can charge $0.001 because the call is cheap: one regex pass, one string join, no upstream fetches, no LLM. The price floor sets a precedent: even the simplest text utility can be paid-callable when the cost is right.

Call it.