AI Control Plane Documentation
Speakeasy AI Control Plane is a platform for securely scaling AI usage. Secure and centrally manage Agents, MCPs, and Skills your whole company can use, with fine-grained permissions, agent identity, security policies, and full observability.
What you can do with Speakeasy
Build MCP servers with external OAuth
GuideGovern a third-party MCP server
GuideGovern agent actions with enterprise identity
GuideShadow MCP allow and block lists
GuidePermission AI with enterprise identity
IdentityTie every agent session to a person, an agent, or a remote identity provider, and revoke access in one place.
Learn moreHost and distribute MCP servers
MCP GatewayRun MCP servers behind one governed endpoint, front several with a gateway, and roll out skills and plugins to every agent platform.
Learn moreEnforce security policies
Security and PolicyScan agent traffic for secrets, sensitive data, and shadow MCP servers, then flag or block on a match.
Learn moreTrack AI usage and costs
ObservabilityAttribute spend, tool calls, and agent sessions to people and teams, and export telemetry to existing backends.
Learn moreAgent Platforms
| Platform | Status |
|---|---|
| Cursor | |
| OpenCode | |
| Mistral | Coming soon |
| Gemini | Coming soon |
| Coming soon | |
| Custom Agents |
Agent-friendly docs
These docs are written for AI agents as much as for people. Every page has a Copy Markdown button and can be handed straight to an assistant with Open in Claude, ChatGPT, or Cursor — or agents can pull the raw sources directly. Learn more in Agent-friendly docs.
Markdown for every page
Append .md to any page URL for the raw markdown — the same content behind each page's Copy Markdown button.
llms.txt
A curated index of the most important pages, written for agents deciding what to read first.
Learn morellms-full.txt
The entire documentation set in a single file, for agents that want everything in one request.
Learn more