Connect your application or coding agent to an Action Gateway session using MCP, the Python SDK, the Go SDK, or the TypeScript SDK.
Action Gateway How-Tospublic
Generated on 26 Oct 2026
Action Gateway connects applications and AI agents to tools through a managed MCP endpoint or SDK, with the accounts and permissions you choose. See What You Can Build for standalone agent examples.
Connect Tools
Connect accounts such as GitHub or Linear to Action Gateway, authorize application users, and restore access when credentials expire.
Add a remote MCP server to Action Gateway with no authentication, an API key, or an OAuth connection.
Use Action Gateway to reach tools, MCP servers, and internal services on your VPC or a connected private network without public exposure.
Use Action Gateway Tool Playground to find and run tools in a model conversation, then inspect calls and results.
Manage Tool Access
Group Action Gateway tools into a versioned Toolbelt, then select that version in a session or environment spec.
Create an Action Gateway session to get an MCP endpoint URL with the tools and permissions your client needs.
Require approval for Action Gateway tool calls and handle the decision through MCP elicitation or a separate approval request.
Create an Action Gateway output view that keeps selected result fields and bind it to a new session.
Configure production agents with selected tools, explicit permissions, preloading, direct Python calls, and focused tool outputs.