MCP.so
Sign In

MCP Agent Server

@fkesheh

Repository

About MCP Agent Server

An innovative Model Context Protocol hub that seamlessly connects your client with intelligent agents, enabling powerful cross-platform communication across multiple MCP ecosystems

#agents#agentic-workflows

Basic information

Category

AI & Agents

License

MIT

Runtime

node

Transports

stdio

Publisher

fkesheh

Config

Add this server to your MCP-compatible client using the configuration below.

{
  "mcpServers": {
    "mcp-agent-server": {
      "command": "node",
      "args": [
        "/full/path/to/mcp-agent-server/dist/index.js"
      ]
    }
  }
}

Tools

No tools detected

We auto-extract tools from the README. The maintainer can list them under a ## Tools heading to populate this section.

Overview

What is MCP Agent Server?

MCP Agent Server is a Model Context Protocol hub that connects MCP-compatible clients (such as Claude Desktop, VS Code, Cursor) with specialized AI agents. It enables users to create, configure, and compose multiple agents for cross-platform communication across various AI ecosystems.

How to use MCP Agent Server?

Install via npx mcp-agent-server, npm install -g mcp-agent-server, or by cloning the repository and running npm install and npm run build. Configure agents using a JSON file (e.g., my-agents-config.json) and connect the server to MCP clients by editing the client’s configuration (e.g., claude_desktop_config.json). Start the server with npx mcp-agent-server --config my-config.json or test an agent with npx mcp-agent-server test-agent --name "Agent Name" --prompt "Prompt".

Key features of MCP Agent Server

  • Create and manage multiple specialized agents
  • Compose agents into master agents for complex workflows
  • Integrate custom tools or use prebuilt MCP servers (e.g., Sequential Thinking, Brave Search)
  • Support for 17+ LLM providers via Vercel AI SDK v5
  • JSON-based configuration for easy agent and tool setup
  • Selective agent exposure to MCP clients

Use cases of MCP Agent Server

  • Build code analysis agents that use project context and sequential thinking
  • Create web search agents with Brave Search for real‑time information retrieval
  • Compose a master agent that delegates tasks to specialized sub‑agents
  • Connect multiple MCP clients (Claude Desktop, VS Code) to a unified agent hub

FAQ from MCP Agent Server

What are the prerequisites for running MCP Agent Server?

Node.js v16 or later is required, along with an API key for your chosen AI model provider (e.g., OpenAI, Anthropic, Google) and any API keys needed by specific MCP servers.

How do I configure agents?

Create a JSON configuration file (e.g., my-agents-config.json) in the project root. Define agents with their name, description, model provider/model, and tools (prebuilt servers, custom MCP servers, or references to other agents). The server automatically loads this file if it exists.

How do I connect MCP Agent Server to Claude Desktop?

Add an entry to your claude_desktop_config.json specifying the command (e.g., npx mcp-agent-server) and configuration file path. Set required API keys as environment variables if needed.

Can I import agents from external files?

Yes, you can use an import tool configuration type to import agents from TypeScript/JavaScript files created with the mcp-ai-agent library, including factory functions or pre‑configured instances.

How do I apply configuration changes after editing?

Run npm run build to rebuild the project, then restart your MCP client. On Windows, fully close the client via Task Manager to ensure changes take effect.

—

(Note: The FAQ section is constructed from explicit statements in the README regarding prerequisites, configuration, client connection, importing agents, and rebuild/restart process.)

Comments

More AI & Agents MCP servers

mcp-hfspace MCP Server 🤗

evalstate

MCP Server to Use HuggingFace spaces, easy configuration and Claude Desktop mode.

387

Sequential Thinking Multi-Agent System (MAS)

FradSer

An advanced sequential thinking process using a Multi-Agent System (MAS) built with the Agno framework and served via MCP.

306

Solon Ai

opensolon

Java AI application development framework (supports LLM-tool,skill; RAG; MCP; Agent-ReAct,Team-Agent). Compatible with java8 ~ java25. It can also be embedded in SpringBoot, jFinal, Vert.x, Quarkus, and other frameworks.

416

Perplexity Ask MCP Server

ppl-ai

The official MCP server implementation for the Perplexity API Platform

2.4K

Model Context Protocol for Unreal Engine

chongdashu

Enable AI assistant clients like Cursor, Windsurf and Claude Desktop to control Unreal Engine through natural language using the Model Context Protocol (MCP).

2.0K