API Documentation

Base URL

https://dariocositore.com/api/

All endpoints return JSON except the generate endpoint, which returns a ZIP archive. CORS is enabled for browser clients. Send Content-Type: application/json on POST requests.

Authentication

GET /api/health is public: it is the deploy check. POST /api/lead is public by design, the site newsletter form posts to it from the browser. The JSON body must carry email and consent: true, plus an optional notes string naming the form (kept as the contact's source property); without consent the server answers 400 Consent required. Resend is the only mailing list behind it: a fresh signup answers 200 with {status:"ok", contact, event}, where event is the user.created emission the welcome cycle listens for, a return visit answers {status:"already", …} without re-emitting, and a rejected request answers error (wording safe to show a visitor) alongside detail (the raw upstream text, for us). POST /api/generate expects a bearer token:

Authorization: Bearer <token>

Missing or wrong credentials return 401 Unauthorized with a WWW-Authenticate header pointing at the OAuth protected resource metadata. When the server has no token configured it returns 500 with Server missing API_TOKEN.

Endpoints

GET /api/health

Service health check. Returns status and version.

curl https://dariocositore.com/api/health

{
  "status": "ok",
  "version": "1.0.0"
}

POST /api/generate

Crawls a public website and returns a ZIP containing a Hermes skill, an OpenClaw package, or both. The crawl stays on the same domain and skips non-HTML assets and noindex pages.

Request body

Example

curl -X POST https://dariocositore.com/api/generate \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $API_TOKEN" \
  -d '{"url":"https://example.com","format":"hermes+openclaw","depth":2,"pages":20}' \
  -o skill.zip

Response

On success: 200 OK with Content-Type: application/zip. The archive includes SKILL.md, a site map, raw page extracts, and OpenClaw package files depending on the requested format.

On failure: JSON error body with a error field. Crawl problems are listed in crawl_errors.

{
  "error": "No pages crawled",
  "crawl_errors": ["HTTP 404 on https://example.com/missing"]
}

Agent Discovery

Machine-readable metadata lives under /.well-known/. Agents and automation tools can read these without scraping HTML.

Workflow Repair Intake

The diagnostics form on this site submits broken workflows to a dedicated triage service:

POST https://elitza-triage.dariozzxx.workers.dev/api/triage
POST https://elitza-triage.dariozzxx.workers.dev/api/complete

Both endpoints accept a JSON payload and a Cloudflare Turnstile token. Use the free diagnostic instead of calling these directly unless you are building an integration.

Errors