NewSkills management: security and system of record

Securely ScaleEnterprise AI.

The AI control plane to understand how AI is used, define what it can access, and enforce security policy across every agent, MCP server, and Skill in your enterprise.

Talk to usStart governingClaude
Trusted byGoogleMistralMoonPayFifth ThirdPlanetScaleLaunchDarkly

01 · The problem

Your security & identity platforms weren't built for AI.

Speakeasy is.

Shadow AI
Without a control plane

AI spreads team by team. Nobody can list every agent, MCP server, and Skill in use

On the control plane

One catalog of approved agents, MCP servers, and Skills, so you understand how AI is used across the company

Sensitive access
Without a control plane

Agents reach production systems and customer data through borrowed credentials and shared service accounts

On the control plane

Every agent acts under its own identity, with access defined by team and role in the identity provider you already run

Enforcement
Without a control plane

Policies live in wiki pages and vary by team. There's no way to enforce them or prove they were followed

On the control plane

Policy is defined once and enforced on every action, with every allow-or-deny decision logged for audit

We built our access controls on IAM. When we rolled out agents, the controls broke down. Speakeasy let us extend our identity management to all our AI usage.

Thierry Dang

Security Operations

MoonPay

02 · The platform

Every Agent Action.
Governed on one control plane.

Every AI agent, MCP server, and Skill in your enterprise runs through one governed path. Understand how each is used, define what it can access, and enforce those rules on every action.

Read the guide: what is an AI control plane?

Govern AI

Understand how AI is used across the company and define what it can access. One catalog of approved tools and Skills, access scoped by team and role through the identity provider you already run, and an audit trail behind every decision.

Approved tools and Skills by team and roleAccess follows your identity providerEvery action in one audit trail

Secure AI

Enforce those rules on every action. Each agent holds only the access its role allows, and every prompt, response, and tool call is checked against policy before it reaches your systems. Threats like prompt injection, PII exposure, and leaked credentials are blocked in flight.

Policy enforced on every actionLeast-privilege access for every agentInjection, PII, and secrets caught in flight

Works with every identity provider

Identity providers were built for people signing in to apps. Speakeasy extends the one you already run to AI agents, so every agent gets its own identity and access follows your existing teams and roles.

OktaMicrosoft Entra IDAuth0WorkOSGoogle WorkspacePing IdentityAny SAML / OIDC

03 · How it works

Works with your existing security and AI stack

Speakeasy deploys through the MDM you already run and sits on the path between your agents and your systems. Identity comes from your identity provider, and every action is checked against policy at the control plane.

01

Deploy through your MDM

Roll out to every device with Jamf, Intune, or the MDM you already run. Coverage is fleet-wide from day one, with no per-team setup.

02

Govern every agent

Claude, ChatGPT, Cursor, and the agents your teams build all route through the same control plane, each acting under its own identity from your identity provider.

03

Enforce policy in flight

Every prompt, response, and tool call is checked in real time. Allowed actions continue to your systems, violations stop at the plane, and every decision lands in the audit trail.

04 · Results

When AI is governed,
adoption grows

Numbers from teams running the control plane in production: full-company rollouts, MCP servers shipped in days, and usage that grows because governance stops being per-project work.

Read the MoonPay story

100%

Employee AI rollout

Fermatrolled out governed AI to the whole company.

60K

agent sessions governed per week

MoonPayenforces policy across every agent.

200%

MCP usage growth

PlanetScalecustomers are now agent-based.

05 · Compliance

Your compliance needs an AI control plane

The future of compliance depends on being able to answer three questions: what AI is in use, what can it access, and who approved each action. Speakeasy's control plane produces the answers.

Speakeasy is SOC 2 Type II, ISO 27001 and CSA STAR AI certified, GDPR compliant and HIPAA ready.

EU AI Act

Logging, human oversight, and risk controls for AI systems in production. The catalog, approval flow, and audit trail supply the operational evidence.

ISO 42001

The AI management system standard. Governed catalogs, scoped access, and decision records map to the controls an auditor asks to see.

NIST AI RMF

Govern, map, measure, manage. Fleet-wide visibility and runtime enforcement cover the functions that need more than policy documents.

Compliance at Speakeasy

The control plane meets the standards it helps you enforce. Speakeasy is SOC 2 Type II audited, ISO 27001 certified, and GDPR compliant, with reports available through the trust center.

Visit trust center

Becoming AI-native meant that we needed an entirely new governance & observability stack. You can't use old tools to cover your team's new way of working.

Shreyas Kumar

Co-founder, Fermat

Fermat

Questions

What is an AI control plane?
An AI control plane is a centralized platform that gives organizations governance and security over every AI agent, MCP server, and Skill across the company. Speakeasy provides a single layer to understand how AI is used, define what it can access, and enforce those policies on every action.
What AI agents does Speakeasy support?
Speakeasy works with Claude, ChatGPT, Cursor, GitHub Copilot, and any AI agent that supports the MCP protocol. A single policy layer governs all of them, so teams can use the tools they prefer without creating security gaps.
How does Speakeasy handle shadow AI?
Speakeasy maintains one catalog of approved agents, MCP servers, and Skills, so security can see how AI is used across the company. Connections to MCP servers outside the catalog are detected and quarantined, so unapproved tools can't quietly reach your systems, and teams can request approval to bring them onto the governed path.
How are AI policies defined?
Policies are written once at the control plane, in language the owning team can read, and scoped to the person or agent, the system, and the action. Access follows teams and roles from your identity provider, reads and writes are treated as different privileges, permissions go down to the individual tool, and a matching deny always wins.
How are policies enforced?
Every prompt, response, and tool call crosses the control plane, where it's checked against policy in real time before it reaches your systems. There is no side channel to bypass, and enforcement is the same whether an employee is driving the agent or it runs autonomously. Every allow-or-deny decision is logged, and violations can alert Slack, PagerDuty, or your SIEM.
Does Speakeasy replace my identity provider?
No. Speakeasy extends the identity provider you already run, including Okta, Microsoft Entra ID, Auth0, and any SAML or OIDC provider, to AI agents. Identity providers were built for people signing in to apps, so agents end up on borrowed credentials and shared service accounts. On the control plane, every agent acts under its own identity, and access follows your existing teams and roles.
How does Speakeasy handle security and compliance?
Each agent holds only the access its role allows, and every prompt, response, and tool call is checked against policy in real time. Threats like prompt injection, PII exposure, and leaked credentials are blocked before they leave your perimeter. The platform is SOC 2 Type II and ISO 27001 certified, with a full audit trail for every session.
How long does it take to deploy?
Most teams are up and running within a day. Speakeasy integrates with existing SSO providers like Okta and Entra ID, so access inherits your current groups and roles with no manual provisioning. Self-hosted and cloud deployment options are available.
What is MCP and why does it matter?
MCP (Model Context Protocol) is an open standard that lets AI agents connect to external tools and data sources. Speakeasy provides a managed platform for deploying, securing, and observing MCP servers at scale, turning any internal API or SaaS tool into a governed AI capability.
Can Speakeasy be self-hosted?
Yes. Speakeasy can be deployed in your own cloud or on-premises for full data residency control. The self-hosted option has no external dependencies and meets the requirements of teams that can't send data to third-party infrastructure.

AI everywhere.

Control here.

Talk to usExplore the AI Control Plane