DomainMan

DomainMan was born from the fact that buying domains manually is painfully slow. To acquire a domain you need: WHOIS lookup, TLD language detection, owner search through company registries, email discovery, and then you still have to write a proper message in the owner's native language. This automates the entire flow.

You enter a domain and it runs WHOIS, RDAP, DNS, SSL certificate, CT log, archive.org, and hosting lookups in parallel — using Promise.allSettled() so if one source fails, the rest keep going. Language is detected from the TLD (50+ country codes), with fallback to registrant country code, filtering out privacy proxy countries (Iceland, Panama, Cayman Islands).

Contact discovery is the serious part: the system scrapes the website, extracts JSON-LD schema, analyzes meta tags, discovers contact pages from robots.txt and sitemaps, and for .hu domains also looks for company registration numbers and tax IDs. Email addresses are generated with pattern-based permutations and validated against SMTP MX records — it even handles Hungarian name order (Kovács János → kovacs.janos@, janos.kovacs@).

There's also a domain graph analysis: 3-level deep mapping of related domains (same IP, reverse DNS, shared nameservers), with connection strength scoring and a 30-second time cap. Infrastructure noise (Cloudflare, Hetzner, AWS, Vercel) gets filtered out.

It generates email templates in 21 languages, sends via Resend, with a reply token system — the owner can respond through the platform without the buyer's personal data being shared. Next.js 16, Drizzle ORM, SQLite.