Hey Product Hunt 👋 We built AgentEyes after realizing there is a new reconnaissance surface that almost nobody is monitoring. Attackers can ask ChatGPT, Claude, Gemini or Perplexity about a company — its employees, infrastructure, login systems, email formats, internal tools, exposed credentials, support procedures and other sensitive information — without ever touching the target network. That means there may be nothing for your SIEM, WAF, IDS or endpoint tools to detect. AgentEyes approaches your company from the attacker's side. Give it a domain and it: 🔎 interrogates multiple public AI models with adversarial queries 🔎 finds potentially exploitable disclosures 🔎 verifies claims against the public attack surface 🔎 identifies exposed credentials, secrets and sensitive intelligence 🔎 connects findings to possible attack paths 🔎 continuously monitors how model knowledge changes There is no agent, integration, DNS change or network access required. The first scan is free, and continuous monitoring starts at $15/month for 3 domains. I’d especially love feedback from security people here: what would you ask an AI model if you were reconning your own company? Alexey Founder, Andeavour

AgentEyes

See what AI reveals about your company before attackers do.

See what AI reveals about your company before attackers do.