OpenOSINT MCP Server — Integrate OSINT into Claude

← Back to Blog

OPENOSINT AS MCP SERVER

Integrate OSINT into Claude Desktop and Claude Code

Published: 2026-05-15

INTRODUCTION

Model Context Protocol (MCP) lets AI assistants call external tools in real time. OpenOSINT ships a full MCP server that exposes all 14 OSINT tools over stdio, letting Claude Desktop or Claude Code run live intelligence lookups on your behalf — no terminal switching required.

When an investigation is underway, the AI decides which tools to chain based on your question, executes them in Python, and feeds the real output back. Tool results cannot be hallucinated because the code runs locally.

WHAT YOU GET

All 14 tools available in Claude's tool palette — see the full tools reference
Investigations happen inline in your chat, no separate terminal
The AI decides which tools to chain based on your question
Results are real: tools execute in Python, not simulated by the model
Session context is preserved across tool calls within one conversation

INSTALLATION

Requires Python 3.10+.

git clone https://github.com/OpenOSINT/OpenOSINT.git
cd OpenOSINT
pip install -e .

Set your API key:

export ANTHROPIC_API_KEY=sk-ant-...

Optional keys for additional tools:

export IP2LOCATION_API_KEY=...   # IP2Location geolocation + VPN detection
export ABUSEIPDB_API_KEY=...     # IP abuse reputation scoring
export HIBP_API_KEY=...          # HaveIBeenPwned breach lookup

CONFIGURE CLAUDE DESKTOP

Edit ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or the equivalent Windows path:

{
  "mcpServers": {
    "openosint": {
      "command": "python",
      "args": ["/absolute/path/to/OpenOSINT/openosint/mcp_server.py"]
    }
  }
}

Restart Claude Desktop. OpenOSINT's tools will appear in the tool palette as search_email, search_ip, search_abuseipdb, etc.

CONFIGURE CLAUDE CODE

From any terminal with Claude Code installed:

claude mcp add openosint python /absolute/path/to/OpenOSINT/openosint/mcp_server.py
claude mcp list

Now inside a Claude Code session you can ask it to investigate targets and it will call the appropriate OSINT tools automatically.

EXAMPLE INVESTIGATION

In Claude Desktop or Claude Code, type a natural-language request:

> Investigate the email suspicious@example.com. Check for linked accounts and breach exposure.

Claude will call tools in sequence:

  → generate_dorks('suspicious@example.com')
  → search_email('suspicious@example.com')
  ✓ Found: GitHub, WordPress, Gravatar

  → search_breach('suspicious@example.com')
  ✓ Found in 1 breach: LinkedIn (2016)

  Summary: The email is registered on GitHub, WordPress, and Gravatar.
  It appeared in the LinkedIn 2016 breach (leaked: email, password hash).

CLI VS MCP

CLI	MCP + Claude
Run one tool at a time	AI chains tools automatically
Output is raw text	Results are interpreted and summarised
You decide the workflow	AI selects tools based on findings
No memory between commands	Session context preserved

For scripting and quick one-off lookups, the CLI remains fastest. For open-ended investigations where you are not sure which tools to use, MCP lets the AI drive.