mcp-client-ref
@ananis25
Basic information
About mcp-client-ref
TUI to test model context protocol servers
Overview
What is mcp-client-ref?
mcp-client-ref is a terminal-based Model Context Protocol (MCP) client built with the Textual TUI library. It lets you connect to MCP servers and interact with them via a language model (defaults to Claude Haiku).
How to use mcp-client-ref?
Install dependencies with uv sync, then run uv run scripts/client.py <mcp_servers_config_path>. The config file is a JSON file that lists your MCP servers in the format used by Claude Desktop. Set ANTHROPIC_API_KEY in a .env file for the default model. You can change the model by editing the client code.
Key features of mcp-client-ref
- Terminal-based TUI interface using the Textual library
- Connects to MCP servers via a JSON config file
- Defaults to Claude Haiku (configurable via code)
- Includes an MCP server inspector for debugging
Use cases of mcp-client-ref
- Debug and test MCP servers with the built‑in inspector
- Interact with MCP servers through a terminal LLM client
- Run a lightweight MCP client without a full IDE
FAQ from mcp-client-ref
What model does mcp-client-ref use by default?
The client uses the Claude Haiku model by default, so you need an ANTHROPIC_API_KEY in your .env file.
Can I use a different model?
Yes, the client uses the magentic library for LLM connections, and you can configure a different model in the client code.
How do I specify which MCP servers to connect to?
Pass the path to a JSON config file as a command‑line argument. The file follows the format used by the Claude Desktop app.
Is there a way to debug MCP servers?
Yes, you can open the MCP server inspector screen from the command palette inside the application.
More Agent Frameworks MCP clients
MCP REST API and CLI Client
rakesh-eltropyUnified MCP Client Library
pietrozullomcp-use is the easiest way to interact with mcp servers with custom agents
LangChain.js MCP Adapters
langchain-ai** THIS REPO HAS MOVED TO
MCPHost 🤖
mark3labsA CLI host application that enables Large Language Models (LLMs) to interact with external tools through the Model Context Protocol (MCP).

CreceRank | Mejor herramienta GEO en Español
CreceRank | Mejor herramienta GEO en EspañolCreceRank es la mejor herramienta GEO en Español. Puedes usar su MCP con esta url: https://crecerank.onrender.com/mcp
Comments