01 / Inspect

Check a site. Draft the three files.

Paste an origin. Agentfile fetches robots.txt and llms.txt from a Worker (not the browser), reports what is present, and writes a sensible llms.txt, a minimal robots.txt, and a starter sitemap.xml. Review before you publish.

Name override (optional)

Description override (optional)

Free: 30 inspect/generate requests per IP per UTC day. Paid key: 600/day. Generated files are drafts.

02 / Status

Result

03 / Generated files

Copy these onto the origin

Suggested /docs, /api, and /blog paths are included only when those URLs respond. Otherwise they are marked as suggestions.

/llms.txt

/robots.txt

/sitemap.xml

04 / API

Same worker, JSON responses

Responses send X-Agentfile-Plan: free or paid. After checkout, Creem emails the license key. Send it as Authorization: Bearer or X-Agentfile-Key for 600 requests per UTC day. Get Agentfile API · Pay in CNY (爱发电). International cards $9/mo via Creem. CNY ¥19/mo via Afdian.

GET	`/v1/check?url=`	Presence and validation
POST	`/v1/generate`	`{ url, name?, description? }` → files
POST	`/v1/keys/verify`	`{ key }` — Creem license check

Paid: Authorization: Bearer <key> or X-Agentfile-Key. Example: curl -sS "ORIGIN/v1/check?url=https://example.com"

05 / Notes

What this does not do

It does not crawl a whole site, invent traffic numbers, or replace a real sitemap generator. Private hosts and loopback addresses are rejected. Fetched bodies are not stored; a per-IP counter is kept in KV for the daily free cap, or per-key for a valid paid license. Publisher: Laurel. Contact: lovlace77777@gmail.com.