OpenOSINT v2.14.0 — AI-powered OSINT agent: 14 tools, interactive REPL, MCP server & web UI
I built @OpenOSINT because every OSINT workflow I tried
was the same story: 5 different tools, manual copy-paste between them,
no memory, no report. I wanted one agent that thinks.

What OpenOSINT does:
You give it a target — email, username, domain, IP, phone number — and the
AI decides which of the 14 built-in tools to run, chains them based on
findings, and writes a structured report.
No prompting required.
Three ways to use it:
→ Interactive AI REPL in your terminal (`openosint`)
→ Browser-based chat UI with streaming output (`openosint web`)
→ MCP server for Claude Code / Claude Desktop (all 14 tools exposed)
What it can find:
Social accounts across 300+ platforms, data breach exposure, WHOIS,
subdomains, IP geolocation + VPN/Tor/proxy detection, abuse reputation,
open ports via Censys, Pastebin dumps, Google dork URLs, phone intelligence.
The AI layer uses Anthropic's native tool use API — hallucination in tool
results is structurally impossible because real tools run and real output
comes back.
Install in 2 lines:
pip install openosintopenosintMIT license. For authorized security research only.
Would love your feedback — especially on the web UI which just launched.
Web Demo:

repo: https://github.com/OpenOSINT/OpenOSINT
site: https://openosint.tech/
Replies