Certiv is a Seattle-based startup founded in June 2025 that emerged from stealth in March 2026 with $4.2 million in pre-seed funding. It provides endpoint AI agent governance for mid-market and large enterprises, particularly those with substantial use of coding agents in software, SaaS, financial services, healthcare technology, and insurance. Its focus is governing AI agents directly on workstations and servers, including shadow agents, custom agents, desktop agents, browser agents, local-model agents, and their Model Context Protocol servers.
The platform combines Scout, a lightweight endpoint policy enforcement point, with Sage, a cloud-hosted policy decision point. Scout intercepts agent model requests, tool calls, shell commands, file operations, and local inference activity; Sage evaluates each action against policy, context, and observed intent before execution. Available responses include allowing and recording activity, blocking it, pausing it for approval, steering the agent to a safer alternative, or redacting secrets.
Certiv’s endpoint architecture addresses activity that gateway, CASB, EDR, and AI-gateway controls may not see, especially local model inference and local tool use. It records agent-session context, analyzes stated, behavioral, demonstrated, and drifting intent, and detects attempts to circumvent denied actions. The platform also tracks token use and spending by user, endpoint, and model.
Privacy Mode obscures compliant agent sessions while exposing policy violations, with justified break-glass access for investigations. Key limitations include dependence on endpoint deployment, lack of coverage for SaaS-embedded and managed-cloud agents, early-stage maturity, limited SIEM integrations, cloud dependence for policy decisions, and continued compliance and European-hosting gaps. Certiv’s central challenge is proving accuracy, scaling customers, and competing against established security platforms adding similar controls.
See All Locations
See All Locations