The new release adds four capabilities to Cequence AI Gateway — AI Discovery, API Registry, LLM Registry, and Skill Registry — and upgrades the company's Agent Personas feature, which binds an agent's permitted models, tools, and API access to a single job description rather than relying on manually constructed policy maps updated for each new use case.
The framing Cequence uses internally is Agentic Zero Trust: nothing an agent touches is implicitly trusted, and every action is scoped to what its defined job requires. Prior solutions have addressed one channel at a time — a prompt filter here, an API gateway there — leaving the surface where MCP, LLM, and API channels intersect ungoverned. Under this release, all three run through a unified identity: the Agent Persona.
AI Discovery surfaces agents, LLM providers, and MCP servers already running across an enterprise by pulling from existing SIEM logs, whether or not they went through a formal approval process. API Registry lets agents call approved APIs without holding the underlying credentials; they authenticate with a single AI Gateway access key. LLM Registry applies the same brokered-credential model to language model calls, adding built-in data loss prevention that inspects prompts and responses for blocked content, including base64-encoded payloads and invisible Unicode characters used to evade filters. Skill Registry gives security teams a governed catalogue of capabilities that can be reused across use cases without re-vetting each time.
The combined effect, Cequence says, is that a finance, HR, or marketing team member can deploy a governed AI agent without waiting on a security review cycle. The Agent Persona defines the boundary; the policy engine enforces it automatically.
All four capabilities are live now as part of AI Gateway for existing Cequence customers.
To stay across the latest in cloud, AI and enterprise tech analysis from Compare the Cloud, subscribe to our weekly newsletter at https://www.comparethecloud.net/newsletter
