$ loading_
Manage local MCP agent connections, tool exposure, and upstream configs centrally.
Copy the install command and let the AI configure it · recommended for beginners
No copy-paste install info for "mcp-hub" yet — see the docs or source repo.
Help me design an mcp-hub configuration so Claude Desktop, Cursor, and another local MCP client can share one local agent, with tool naming, visibility, and upstream sources managed from a remote control plane.
A configuration plan, component layout, and setup guidance for sharing one local agent across multiple clients.
I want to use mcp-hub to set different tool exposure policies for different MCP clients: developer clients can access all tools, while office clients can only access document and search tools. Provide configuration guidance and naming conventions.
A configuration approach for per-client tool visibility and clear tool naming recommendations.
Explain how to use mcp-hub to combine local stdio tools, remote configuration, and multiple upstream MCP sources into one unified entry point, and list key considerations for permissions, reliability, and maintenance.
An overview of the unified access architecture plus a checklist of deployment risks and best practices.
Use one gateway to lazily spawn and proxy multiple MCP servers.
Manage and monitor MCP servers centrally with dynamic configuration and control.
Create, manage, and compose AI agents for MCP-compatible clients and tools.
Aggregate multiple MCP servers into a few fixed tools for AI apps.
Aggregate local MCP servers into one endpoint with runtime control and hot reload.
Manage multiple coding agents centrally with parallel orchestration across projects.