Certiv
Certiv builds endpoint software that discovers AI agents running on employee devices and checks their model requests and tool calls against policy before they execute.
About Certiv
Certiv is a Seattle-based startup selling what it calls AI agent assurance for endpoints.
A lightweight agent installed on Mac, Windows and Linux machines, as well as containers, finds the AI agents employees run, reads the context they operate in, and decides whether each model request or tool call may proceed before it takes effect.
A separate judgment layer evaluates what the agent is trying to do, while a management console holds policies and gives security teams visibility.
The company says the goal is to let enterprises adopt coding, operations and knowledge-work agents without giving them unchecked access to live systems. Certiv emerged from stealth in March 2026 with a pre-seed round.
Investors
Work at Certiv? Link to this listing
Optional badge for your site or README. It changes nothing about the listing, its order or its review. Other formats and a dark version: press kit.
<a href="https://securitystartups.com/companies/certiv"><img src="https://securitystartups.com/badge/listed.svg" alt="Listed on Security Startups" width="184" height="44"></a>