Unified Identity and Access Infrastructure for AI Agents

Keycard is developing identity infrastructure for environments where AI agents interact autonomously with systems and services. The company focuses on secure identity management for AI agents and platforms built on the Model Context Protocol (MCP). Its offering centers on providing transparent access controls, contextual policy enforcement, and adaptive authorization frameworks tailored to developers building agent-based systems. The platform is designed to manage identity and access flows that traditional human-centric IAM tools do not address.

Keycard targets challenges introduced by autonomous agents that require time-bounded, resource-specific, and revocable credentials across distributed computing environments. The system provides mechanisms to govern how AI agents authenticate, obtain permissions, and interact with organizational resources.

Founded in 2025 in the United States, Keycard announced both a seed round ($8 million, co-led by Andreessen Horowitz and boldstart ventures) and Series A ($30 million, led by Acrew Capital) in October 2025. Angel investors include Matias Woloski (co-founder and former CTO at Auth0), Emilio Escobar (CISO at Datadog), Ryan Carlson (president at Chainguard), Ian Andrews (CRO at Groq) and Karl McGuinness (former Chief Product Architect at Okta).