Vibeteam

An MCP (Model Context Protocol) server that intelligently orchestrates sub-agents based on query analysis and project context.

Features

• Query Analysis: Understands user intent (build, debug, feature, refactor, research, etc.)
• Project Detection: Automatically identifies project type (backend, frontend, fullstack, library)
• Smart Agent Orchestration: Deploys specialized agents based on project context and query intent
• Parallel Execution: Runs multiple agents concurrently for faster results
• Tool Detection: Automatically adapts to Cursor, Windsurf, VS Code, Claude Desktop, etc.
• Multi-Agent Support:
  • Backend Specialist
  • Frontend Specialist
  • Web Researcher
  • Code Analyzer
  • Test Engineer
  • DevOps Specialist

Installation

Via uvx (Recommended)

Once published to PyPI, you can install and run directly with uvx:

uvx vibeteam

Via pip

pip install vibeteam

From Source

git clone https://github.com/XiaoConstantine/vibeteam.git
cd vibeteam
uv sync
uv run vibeteam

MCP Configuration

Add to your Claude Desktop or MCP client configuration:

{
  "servers": {
    "vibeteam": {
      "command": "uvx",
      "args": ["vibeteam"]
    }
  }
}

Or if installed via pip:

{
  "servers": {
    "vibeteam": {
      "command": "vibeteam"
    }
  }
}

Usage

As MCP Server

The server runs as an MCP server and exposes tools for Claude or other MCP clients:

# Run with uvx
uvx vibeteam

# Or if installed
vibeteam

Testing

uv run python test_server.py

Tools Available

1. analyze_and_orchestrate: Main orchestration tool that analyzes query and deploys agents
2. analyze_project: Analyzes project structure and technology stack
3. analyze_query: Determines query intent and complexity

How It Works

1. Project Analysis: Scans the current working directory to understand project structure
2. Query Understanding: Analyzes user query to determine intent and complexity
3. Agent Planning: Based on project type and query intent, plans appropriate agents
4. Orchestration: Executes agents in priority order and collects results
5. Summary: Provides consolidated output from all agents

Architecture

The orchestrator uses a sophisticated analysis system to:

• Detect backend (Python, Go, Java, etc.) and frontend (React, Vue, Angular) technologies
• Understand query complexity and required expertise
• Deploy agents with appropriate priority levels
• Coordinate multi-agent workflows for complex tasks