Bot & Crawler Detection
Automated crawlers, SEO spiders, AI scrapers, and uptime checkers often pollute standard analytics data. VeloStats filters and categorizes this traffic server-side.
Categorization Engine
Every incoming request passes through an autonomous inspection pipeline:
- User-Agent Fingerprinting: Known search spiders (Googlebot, Bingbot), social crawlers (Twitterbot, LinkedInBot, WhatsApp), and AI crawlers (GPTBot, ClaudeBot, Perplexity) are detected and classified.
- Datacenter ASN Matching: IP addresses originating from major hosting providers (AWS, Hetzner, DigitalOcean, OVH) are flagged using local MaxMind GeoLite2 ASN databases.
- Behavioral Traps & Diagnostics: Headless Chrome, webdriver flags, and impossible interaction timings trigger suspicious bot classifications.
Verdict Types
| Category | Description | Counted in Humans? |
|---|---|---|
search_engine |
Legitimate indexing crawlers (Google, Bing). | No |
ai_crawler |
LLM scrapers (OpenAI, Anthropic, Common Crawl). | No |
social_media |
Link-preview generators (Slack, Discord, Apple iMessage). | No |
datacenter |
Server-originating requests without human interaction. | No |
human_verified |
Verified real human session (passed heartbeat checks). | Yes |