Bot-Detect API — Overenie crawlerov vyhľadávačov | YEB

#What Bot-Detect solves

This API verifies whether an IP truly belongs to an official crawler (e.g., Googlebot, Bingbot) instead of someone claiming to be a bot via User-Agent. It’s designed for abuse prevention, SEO safety (don’t block real bots), and traffic classification (bill bots differently, slow them down, or allow them).

#Endpoints and when to use them

#POST /v1/bot/detect — Auto-detect

• Best for: General checks when you don’t know the vendor in advance.
• How it works: If a ua is provided, we try to infer a vendor from it; otherwise we test your IP against all supported vendors’ ranges.
• Typical use: Single integration point for all traffic (edge/middleware).

#POST /v1/bot/detect/{vendor} — Vendor-specific

• Best for: When routing already knows a suspected vendor (e.g., URLs dedicated to Google).
• Vendors: google, bing, duck, qwant, meta, yandex, seznam, openai.
• Note: UA is informative; the final decision is IP-based (with optional reverse-DNS/ASN checks).

#Quick start

curl -X POST "https://api.yeb.to/v1/bot/detect" \
  -H "Authorization: Bearer <YOUR_API_KEY>" \
  -H "Content-Type: application/json" \
  -d '{ "ip": "66.249.66.1", "ua": "Mozilla/5.0 (compatible; Googlebot/2.1; +http://www.google.com/bot.html)" }'

# Vendor-specific (Google)
curl -X POST "https://api.yeb.to/v1/bot/detect/google" \
  -H "Authorization: Bearer <YOUR_API_KEY>" \
  -H "Content-Type: application/json" \
  -d '{ "ip": "66.249.66.1" }'

// JS Fetch example
fetch('https://api.yeb.to/v1/bot/detect/bing', {
  method: 'POST',
  headers: {
    'Authorization': 'Bearer <YOUR_API_KEY>',
    'Content-Type': 'application/json'
  },
  body: JSON.stringify({ ip: '40.77.167.129' })
})
.then(r => r.json())
.then(console.log)
.catch(console.error);

#Parameters that actually matter

#Reading and acting on responses

You mostly care about result.ok, result.vendor, and result.reason.

{
  "result": {
    "vendor": "google",
    "ok": true,
    "reason": "ip_match",                // or "ip_and_ua_match", etc.
    "ua_present": true,                  // UA was sent explicitly as param
    "ua_source": "param",                // "param" | "header" | null
    "ua_match": true,                    // UA pattern fit the vendor
    "ip_match": true,                    // IP fell inside vendor ranges
    "dns_verified": false,               // set true if verify_rdns passed
    "rdns_checked": false,
    "asn_verified": false,
    "asn_checked": false,
    "cidr_empty": false,                 // true if no ranges were available
    "ip_kind": "search_bot",             // Google only: search_bot | special_crawler | user_triggered_google | user_triggered_user | cdn_proxy | unknown
    "ip_kind_source": "json",            // "json" | "dns_ptr" | null
    "ptr": "crawl-66-249-66-1.googlebot.com"
  }
}

#Typical reason values

• ip_and_ua_match — Best case: UA matched the vendor pattern and IP belongs to the vendor.
• ip_match — Good: IP belongs to vendor ranges; UA was missing or didn’t match (still OK to treat as vendor).
• ip_match_but_ua_not_matched — Likely a legitimate bot with a non-standard UA; don’t block.
• ua_not_matched — UA said “I’m X”, but we couldn’t confirm it; check ip_match.
• ip_not_in_vendor_ranges — Treat as non-vendor; consider bot throttling or block rules.

#Recommended actions

• ok = true → Allow, skip WAF challenges, crawl-budget friendly rate-limit.
• ok = false and UA claims vendor → 403 or serve simplified page; log for audit.
• Need absolute certainty → Call with verify_rdns=true (+ strict_rdns=true if you want hard fail).

#Troubleshooting & field notes

1. “UA says Googlebot but ok=false” → You saw a spoofed UA. Key is ip_match. Consider returning 403 or a lightweight page.
2. Rapid vendor changes → Prefer vendor JSON ranges (handled automatically). If you need hard proof, enable verify_rdns.
3. IPv6 traffic → Fully supported. If your edge strips IPv6, fix that first.
4. False negatives for Meta → Use verify_asn=true (and asn=32934 if relevant) for stronger confirmation.
5. Rate-limits → Respect 429 with exponential backoff. Keep request IDs in your logs.

#API Changelog