Agent Security
Checks every prompt, response and tool call of AI agents against policy in flight, detecting shadow AI and data leaks and blocking risky actions across agents, MCP servers and Skills.
Find alternatives to Agent Security
What it does
- Guardrails and safety
- Data security
Sources
- Pricing
-
Pricing page: pay-as-you-go usage meters and an Enterprise annual contract, both for the one control plane; no separate rate for Agent Security.
- Description
-
Official documentation · 16 Nov 2026
Product page text fetched 2026-11-16 with curl; page quote: "Every prompt, response, and tool call is checked against policy in flight". The Products menu lists it under AI Control Plane.
- What it does
-
Official documentation · 16 Nov 2026
Product page text fetched 2026-11-16 with curl; page quote: "Every prompt, response, and tool call is checked against policy in flight". The Products menu lists it under AI Control Plane.
- Sold within
-
Pricing page: "Both plans carry the same control plane: the same connectors, the same OAuth, provisioning and RBAC controls, and the same logs, tracing and analytics." The page prices the control plane's usage only and no line names Agent Security, so it is part of the control plane's plans.
- Name
-
Official documentation · 16 Nov 2026
Product page text fetched 2026-11-16 with curl; page quote: "Every prompt, response, and tool call is checked against policy in flight". The Products menu lists it under AI Control Plane.
- URL
-
Official documentation · 16 Nov 2026
Product page text fetched 2026-11-16 with curl; page quote: "Every prompt, response, and tool call is checked against policy in flight". The Products menu lists it under AI Control Plane.
- Deployment
-
Official documentation · 16 Nov 2026
Product page text fetched 2026-11-16 with curl; page quote: "Every prompt, response, and tool call is checked against policy in flight". The Products menu lists it under AI Control Plane.
- Interfaces
-
Official documentation · 16 Nov 2026
Product page text fetched 2026-11-16 with curl; page quote: "Every prompt, response, and tool call is checked against policy in flight". The Products menu lists it under AI Control Plane.
Also from Speakeasy
-
MCP Gateway Platform
MCP Gateway gives enterprises one governed endpoint for every MCP server, with identity-provider sign-in, tool-level access control and an audit log of every tool call.
-
AI Control Plane Platform
An AI control plane for every agent in an enterprise, built around an MCP gateway, with agent identity from the company's identity provider, security checks on every prompt and tool call, and an audit trail of sessions, tokens and spend.
-
Agent Identity Platform
Gives every AI agent a persistent identity: sessions sign in through the company's identity provider, tool calls are authorised against readable grants with short-lived task-scoped credentials, and each decision is logged against a name.
-
Agent Observability Platform
One audit trail of every agent session, tool call, token and dollar attributed to a named person in real time, with spend attributed to teams and budgets and exportable to a SIEM.
Open Speakeasy in the directory
Something wrong here? Send a correction — quote this product id: speakeasy-agent-security.