mcpserver.so LogoMCPFly
Submit
MCPHost logo

MCPHost

by mark3labs

AI/Development ToolsCLIClaudeOllamaGeminiOpenAITooling

MCPHost is a CLI host application that enables Large Language Models (LLMs) to interact with external tools through the Model Context Protocol (MCP). It currently supports Claude 3.5 Sonnet and Ollama models, along with Google Gemini and OpenAI-compatible models.

View on GitHub

Last updated: N/A

What is MCPHost?

MCPHost acts as a host in the MCP client-server architecture, allowing LLMs to access external tools and data sources, maintain consistent context, and execute commands safely. It provides a command-line interface for interacting with various AI models through MCP servers.

How to use MCPHost?

Install MCPHost using go install github.com/mark3labs/mcphost@latest. Configure MCP servers in the ~/.mcp.json file or specify a custom location using the --config flag. Use the CLI with flags to select models (e.g., mcphost -m ollama:qwen2.5:3b) and interact with tools. Utilize interactive commands like /help, /tools, /servers, /history, and /quit during conversations.

Key features of MCPHost

  • Interactive conversations with support models

  • Support for multiple concurrent MCP servers

  • Dynamic tool discovery and integration

  • Tool calling capabilities for both model types

  • Configurable MCP server locations and arguments

  • Consistent command interface across model types

  • Configurable message history window for context management

Use cases of MCPHost

  • Interacting with external databases through LLMs

  • Automating tasks using LLMs and external tools

  • Building custom LLM-powered applications

  • Integrating LLMs with existing systems via MCP servers

FAQ from MCPHost

What is MCP?

MCP stands for Model Context Protocol, a standard for LLMs to interact with external tools.

What models are supported?

Currently supports Claude 3.5 Sonnet, Ollama-compatible models, Google Gemini models, and OpenAI-compatible models.

How do I configure MCP servers?

MCP servers are configured in the ~/.mcp.json file, specifying the command and arguments for each server.

How do I specify which model to use?

Use the -m or --model flag followed by the model name in the format provider:model (e.g., ollama:qwen2.5:3b).

Where can I find MCP-compatible servers?

See the MCP Servers Repository at https://github.com/modelcontextprotocol/servers for examples and reference implementations.