Mindgard is a deep-tech startup specializing in AI security. They provide an automated “AI red-teaming / security testing” platform that helps organizations secure their AI models and agents across the entire lifecycle — detecting vulnerabilities like prompt injection, agentic manipulation, and other runtime threats. They support models including LLMs, multi-modal AI (image, audio, etc.), and integrate into existing CI/CD and software development workflows

Mindgard was spun out from research at a university lab (originally from Lancast