McpManager FAQs

Question 1

What is McpManager and what problem does it solve?

Accepted Answer

McpManager is a self-hosted proxy and aggregation platform for Model Context Protocol (MCP) servers. It solves the problem of configuring multiple MCP servers individually in your AI clients by providing a single, unified endpoint to access all aggregated tools, simplifying management and access for AI agents like Claude, Cursor, and Windsurf.

Question 2

How does McpManager simplify access for AI clients?

Accepted Answer

Instead of connecting your AI clients to several individual MCP servers, you connect them once to McpManager's unified endpoint. This allows your clients to access all tools from all registered upstream MCP servers through a single connection, streamlining configuration and interaction.

Question 3

Can I customize the tools exposed through McpManager?

Accepted Answer

Yes, McpManager offers extensive tool customization. You can override tool names, descriptions, and parameter schemas. It also allows you to enable or disable individual tools per server or namespace, ensuring different tool sets can be exposed as needed.

Question 4

What kind of upstream servers and protocols does McpManager support?

Accepted Answer

McpManager supports multiple transport types for upstream servers, including HTTP/SSE for remote MCP servers, Stdio for local CLI-based tools, and automated conversion from OpenAPI 3.x specifications to MCP tools. Each transport supports its own authentication methods.

Question 5

How does McpManager handle security and access control?

Accepted Answer

McpManager includes robust API Key Management, allowing you to generate scoped API keys with optional namespace restrictions. These keys are used as Bearer tokens for authenticating MCP client connections, with per-request tracking for auditing purposes. It also supports multi-user management with claims-based authorization.

McpManager

McpManager

Key Features

Use Cases

Key Features

Use Cases