Ultra Security, a developer of security tooling for autonomous AI agents, has emerged from stealth with $10m in seed capital, betting that firms can no longer afford to discover what their agents did only after the damage is done.
The round drew backing from Ludlow Ventures, Hidden Capital, Entrée Capital, Detroit Venture Partners, firedrop and FDVC. Angel investors drawn from the security and SaaS sectors also took part.
The company positions its platform as a checkpoint between AI agents and the MCP servers and connectors those agents rely on. Each prompt, data retrieval and tool invocation passes through Ultra, giving organisations live oversight and control of agent behaviour rather than a retrospective view.
Ultra argues that the timing is pressing. Agents have evolved from simple chatbots into software with genuine access to corporate systems, including email and code, and MCP is fast becoming the default route through which they reach tools, APIs and sensitive data. According to the firm, adoption is outpacing the monitoring tools designed to keep watch.
Web application firewalls and API gateways were designed to track exchanges between humans and systems, while cloud posture products assess configuration only at fixed moments. Ultra says none of these can observe agent activity inside a live session, leaving most teams unable to say what their agents are doing, when, and which data they have touched.
The platform combines four capabilities. Its guardrails convert plain-language policies into real-time enforcement, blocking credential access, redacting PII, protecting PHI, limiting abuse and preventing data moving between connectors, with rules applicable to an entire organisation, a team or one device.
Governance controls determine which agents, connectors and tools each user or team may access. Observability features log and trace every agent-to-tool interaction, with AI-native anomaly detection raising alerts. A universal connectivity layer allows a single setup to serve every agent and connector, independent of vendor or model.
Ultra was founded in 2025 by Chase Lee, who previously founded and led Trustpage before its acquisition by Vanta. Lee observed security gaps widening as agents linked into the wider technology stack, while existing products had not been designed for software operating independently.
During its year in stealth, the team collaborated with customers and design partners to simplify agent security and AI governance for organisations regardless of scale, security maturity or AI adoption.
The company stresses that it was designed for agentic security from inception rather than retrofitted by a legacy vendor or an orchestration provider. It runs locally on the user’s machine by default, keeping tool calls, payloads and audit records on-device, with a remote-hosted alternative also offered.
Drawing on a founding team from Vanta and Secureframe, Ultra holds SOC 2 Type II, ISO 27001 and ISO 42001 certifications and is AARM Extended conformant ahead of launch. It aligns with NIST AI RMF and the Agentic Trust Controls (ATC), has completed a FedRAMP 20x Class B audit with a 3PAO, and is pursuing FedRAMP certification.
Copyright © 2026 FinTech Global
Investors
The following investor(s) were tagged in this article.
Leave a Reply