Deep Eye is an open-source AI-driven security assessment tool that orchestrates multiple AI providers to generate intelligent test payloads and scan targets for vulnerabilities. It coordinates models from OpenAI, Claude, Grok, Gemini, Ollama, Groq, Mistral, OpenRouter, LiteLLM, and LM Studio, and produces professional reports with compliance mapping. The project has earned more than 2,300 stars on GitHub and was still actively updated in October 2026.
Traditional vulnerability scanners rely on fixed signature databases, which miss novel attack surfaces and generate noisy results. Deep Eye takes an AI-orchestrated approach: language models help craft context-aware payloads, reason about target responses, and prioritize findings across more than forty-five vulnerability types. Because it can route work across many model providers โ including fully local models through Ollama or LM Studio โ assessments can be tuned for cost, speed, or privacy, and sensitive engagements can stay on local infrastructure. Scan output is organized into professional reports that map findings to compliance requirements, which makes results easier to hand to clients or engineering teams for remediation. Its modular provider layer also means newly released models can be plugged in as they appear, keeping the assessment workflow current without redesigning the tool.
Deep Eye is free and open source. If you connect commercial AI providers, you pay their standard API rates; using local models through Ollama or LM Studio keeps assessments entirely cost-free.
Deep Eye is for penetration testers, red teams, and security engineers who want AI assistance in vulnerability discovery and reporting. As with any scanning tool, it should only be used against systems you own or are explicitly authorized to test.