The autonomous security agent that thinks like an attacker. Multi-agent AI with persistent memory, 20+ LLM providers, and 18 built-in pentest tools.
Latest Release v0.2.4 · MIT License · Cross-Platform
Help us build the future of autonomous security testing.
Security researcher and full-stack developer passionate about offensive security and AI.
Security Researcher · Full-Stack Developer · AI Enthusiast
Building autonomous security testing tools to help ethical hackers and security researchers work smarter. Focused on AI-powered offensive security, multi-agent architectures, and making pentesting accessible to everyone.
Point it at a target and let the AI do the rest.
Give PentestCode a target and an objective. It understands natural language.
AI agents plan the attack, dispatch specialists, and coordinate in parallel.
Get a complete report with evidence, attack paths, and remediation advice.
Everything you need for thorough, autonomous penetration testing.
Claude, GPT-4, Gemini and 20+ providers make intelligent security decisions.
Learn more →Remembers hosts, services, credentials, and attack paths across sessions.
Learn more →19 curated packs covering every attack surface and pentest methodology.
Learn more →Built-in parsers for nmap, nuclei, sqlmap, bloodhound, gobuster, and more.
Learn more →Cost-based path finding through relationship graphs with graph algorithms.
Learn more →13 specialized agents working in parallel for faster, deeper testing.
Learn more →A strategist-coordinator dispatches specialist agents in parallel.
Watch PentestCode autonomously test a target.
See how it compares to traditional approaches.
| Feature | PentestCode | Manual Testing | Other AI Tools |
|---|---|---|---|
| Speed | ★★★★★ | ★★ | ★★★ |
| Automation | ✓ Full | ✗ None | ✓ Partial |
| Multi-Agent | ✓ 13 agents | ✗ | ✗ |
| Persistent Memory | ✓ Full state | ✗ | ✗ |
| AI Planning | ✓ Graph-based | ✗ | ✓ Basic |
| LLM Providers | 20+ | N/A | 1-3 |
| Cost | Free (MIT) | High (labor) | $$ |
"PentestCode saved me hours on my last engagement. The multi-agent architecture found things I would have missed."
"The persistent memory is a game-changer. I can pause a session, come back days later, and pick up exactly where I left off."
"Finally, an AI tool that actually understands offensive security. The knowledge packs are incredibly well-researched."
Available for all major platforms.
| Platform | Architecture | Download | Size |
|---|---|---|---|
| Linux | x64 | pentestcode-linux-x64.tar.gz | 47.5 MB |
| Linux | ARM64 | pentestcode-linux-arm64.tar.gz | 47.1 MB |
| macOS | ARM64 (Apple Silicon) | pentestcode-darwin-arm64.zip | 32.2 MB |
| macOS | x64 (Intel) | pentestcode-darwin-x64.zip | 34.7 MB |
| Windows | x64 | pentestcode-windows-x64.exe | 129.6 MB |
Join security researchers who trust PentestCode for autonomous AI-powered testing.