Skip to content

OpenOSINT

AI-powered OSINT agent with interactive REPL, MCP server, and CLI. 20 tools. Works with Claude, GPT-4, or local models. For authorized security research only.

Install

$claude mcp add openosint -- uvx --from openosint openosint-mcp

Install command taken from the project's README or registry entry. Always review code before running it.

mcp-name: io.github.OpenOSINT/openosint

OpenOSINT An OSINT (Open Source Intelligence) agent for security researchers and analysts: 21 investigation tools behind a natural-language interface, plus an MCP (Model Context Protocol) server so any MCP-compatible AI client can drive them directly. Use it as a REPL, CLI, MCP server, or browser Web UI. Tool results come from real executions, and every finding is shown with the tool call that produced it. AI summaries are leads to verify, not proof.

Hosted version — OpenOSINT Monitor. Weekly external checks of your domain's email authentication (SPF/DKIM/DMARC), TLS, security headers and lookalike domains. Ownership-verified; free for one domain. Prefer to run it yourself? Everything below is free and MIT-licensed.

Self-host (free, MIT)

pip install openosint
# Interactive AI REPL (default)
openosint

# Web interface
openosint web

# Direct tool (no AI)
openosint email target@example.com

Try the live demo → — bring your own Anthropic / OpenRouter / Ollama key, no signup.

Usage

Start the REPL and investigate any target — the agent decides which tools to run and chains them on findings:

openosint > investigate target@example.com

  -> generate_dorks('target@example.com')
  -> search_breach('target@example.com')
  Found in 2 breaches: LinkedIn (2016), Adobe (2013)

  -> search_github('johndoe99')   <- pivoted from breach findings
  Found: profile, 14 repos, commit-discovered emails

  Report saved -> reports/2026-05-11_14-32-11_report.md

Features

Capability Details
AI tool chaining The agent selects and chains tools based on findings; describe the target in plain language
21 modular tools Email, username, breach, WHOIS, RDAP, IP, subdomain, dorks, paste, phone, Shodan, VirusTotal, Censys, IP2Location, AbuseIPDB, GitHub, DNS, GDELT news geo, live dork search, URL scraping, SERP footprint
Three AI backends Anthropic Claude (default), local Ollama, or any OpenAI-compatible endpoint (LiteLLM, vLLM, LM Studio, ...)
Native MCP server Passive tools (20) exposed by default, all 25 with --allow-active (Passive by default); runs with uvx, no separate install
Parallel execution --parallel runs complementary tools concurrently via asyncio.gather()
Reports PDF + Markdown auto-saved after every investigation (reportlab optional)
Session history All REPL sessions saved to ~/.openosint/history/; browse with openosint history
Web UI Browser-based AI chat with streaming output, tool cards, light/dark theme

Investigation results land live on the globe — click a point to pivot. Watch the sharper MP4.

Deterministic, synthetic-data demo — every entity shown is fictional; regenerate with demo/web_demo.py. The entities are seeded at the statement layer, not produced by today's mappers — see demo/README.md.


Legal Disclaimer: OpenOSINT is intended for legal and authorized use only. Users are solely responsible for ensuring their use complies with all applicable laws and regulations. The authors accept no liability for misuse. See DISCLAIMER.md.

Tools

Similar plugs

The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.

Build Agentic workflows, RAG pipelines, with rich AI model and tool support on one collaborative workspace. Deploy on cloud, VPC, or self-hosted, so teams move from prototype…

A cross-platform desktop All-in-One assistant for Claude Code, Codex, OpenCode, OpenClaw, Grok Build & Hermes Agent. Only official website: ccswitch.io