JS

jeffvestal/simple-mcp-client

Developer tools

このページは AI により自動翻訳されています。技術的な詳細は原文をご確認ください。 原文を見る

1 stars 6 forks 品質 62 トレンド 13Apache-2.0

Simple chat App that connects to one or more MCP servers allowing the user to explore how chatting with various MCP tools works

概要

A full-stack chat application that connects to LLM providers and utilizes tools from MCP (Model Context Protocol) servers. Features a clean, modern interface with comprehensive tool execution and visualization capabilities. This application is with robust error handling: ✅ : End-to-end tool calling with LLM follow-up responses ✅ : Automatic parameter correction for any MCP server validation errors ✅ : Comprehensive tool failure recovery and OpenAI API compliance ✅ : Collapsible tool call displays with real-time status indicators ✅ : OpenAI, Gemini, and Bedrock compatible APIs ✅ : Remote and local MCP server support with process management ✅ : Live connection status and server health monitoring ✅ : Persistent theme switching with clean design ✅ : Clear chat functionality with message history ✅ : Proper OpenAI API compliance for multi-turn tool conversations ✅ : Intelligent retry logic for validation errors with parameter correction - : Clean UI with markdown support and...

README

以下は原文 README の抜粋です。原文を見る。

Simple MCP Client

A full-stack chat application that connects to LLM providers and utilizes tools from MCP (Model Context Protocol) servers. Features a clean, modern interface with comprehensive tool execution and visualization capabilities.

Current State

This application is fully functional and production-ready with robust error handling:

✅ Complete Tool Execution Pipeline: End-to-end tool calling with LLM follow-up responses
✅ Smart Error Recovery: Automatic parameter correction for any MCP server validation errors
✅ Robust Error Handling: Comprehensive tool failure recovery and OpenAI API compliance
✅ Professional UI: Collapsible tool call displays with real-time status indicators
✅ Multi-LLM Support: OpenAI, Gemini, and Bedrock compatible APIs
✅ MCP Server Management: Remote and local MCP server support with process management
✅ Real-time Status: Live connection status and server health monitoring
✅ Dark/Light Mode: Persistent theme switching with clean design
✅ Chat Management: Clear chat functionality with message history
✅ Conversation History: Proper OpenAI API compliance for multi-turn tool conversations
✅ Retry Mechanisms: Intelligent retry logic for validation errors with parameter correction

Features

  • Modern Chat Interface: Clean UI with markdown support and user/assistant message bubbles
  • Complete Tool Execution: Tools are executed with results fed back to LLM for natural responses
  • Tool Visualization: Expandable tool call displays showing request/response details with status indicators
  • Advanced Error Handling: Automatic recovery from tool failures with intelligent parameter correction
  • OpenAI API Compliance: Proper conversation history management for multi-turn tool interactions
  • MCP Server Integration: Connect and manage multiple MCP servers with automatic tool discovery
  • Local & Remote Servers: Support for both HTTP-based and local process-based MCP servers
  • Process Management: Automatic server startup/shutdown with health monitoring
  • Multi-LLM Support: OpenAI, Gemini, and Bedrock compatible APIs
  • Real-time Updates: Live connection status and server management
  • Responsive Design: Works on desktop and mobile devices
  • Universal MCP Compatibility: Works with any MCP server through generic error parsing

Tech Stack

  • Frontend: React + TypeScript + Vite
  • UI Library: Shadcn/ui + Tailwind CSS + Lucide Icons
  • State Management: Zustand
  • Backend: FastAPI + Python 3.8+
  • Database: SQLite (file-based)
  • Protocol: MCP over JSON-RPC 2.0 (HTTP and stdio)
  • Markdown: ReactMarkdown for rich message rendering

Quick Start

Run the setup script from the project root:

GitHub で README 全文を見る

インストール

This server does not publish a one-line install command.

Open the repository installation guide MCP 初心者?mcpskill.org のガイドから始めましょう →