METATRON

METATRON is an open-source, AI-powered penetration testing assistant that runs entirely on your local machine. Published on GitHub, it has attracted about 4,230 stars and was updated in 2026. Built for Parrot OS and other Linux systems, it pairs a locally running large language model with classic reconnaissance tools, so security testing can be done offline — no cloud services, no API keys, and no subscriptions sending your scan data to anyone else.

Overview

You give METATRON a target IP address or domain. It runs real recon utilities — nmap, whois, whatweb, curl, dig, and nikto — then feeds the combined results to a local model served through Ollama. The AI analyzes the findings, identifies likely vulnerabilities, suggests exploits, and recommends concrete fixes. Everything is stored in a MariaDB database with full scan history across linked tables, so you can revisit, edit, or delete past results and export clean reports for clients or audits. Because every scan is logged end to end, assessments stay reproducible long after they were run.

Key Features

• Fully local AI analysis through Ollama — works completely offline
• Automated reconnaissance with nmap, whois, whatweb, dig, and nikto
• AI-generated vulnerability analysis, exploit suggestions, and remediation advice
• Built-in web and CVE lookup without requiring any API keys
• MariaDB-backed scan history with editing and report export
• Agentic loop that can request additional tool runs during analysis

💳 Pricing

METATRON is completely free and open source under the MIT license. Because the AI runs locally through Ollama, there are no API bills or subscription costs — you only need a Linux machine capable of running a local model.

Who Is It For

METATRON is for penetration testers, security students, and system administrators who want AI assistance during assessments while keeping target data strictly on their own machines — especially in labs, restricted networks, or engagements where cloud AI services are not allowed.