MCP Agentic AI Server

# MCP Server Project Architecture This project implements a robust MCP (Multi-Component Platform) system, supporting both custom and public-facing deployments with integrated monitoring and analytics. ## Architecture Diagram ```mermaid graph TB subgraph "MCP Server Project" style MCP Server Project fill:#1a1a1a,stroke:#ffffff,stroke-width:0px subgraph "Custom MCP Implementation" style Custom MCP Implementation fill:#6a3093,stroke:#a8a8a8,stroke-width:2px CM[Custom MCP Server] MC[MCP Controller] CT[Custom Tools] CM --> MC MC --> CT end subgraph "Public MCP Implementation" style Public MCP Implementation fill:#a8dadc,stroke:#a8a8a8,stroke-width:2px PM[Public MCP Server] GC[Gemini Client] AC[Agent Config] PM --> GC PM --> AC end subgraph "Frontend" style Frontend fill:#00b894,stroke:#a8a8a8,stroke-width:2px SD[Streamlit Dashboard] SD -->|Stats API| CM SD -->|Stats API| PM end subgraph "API Endpoints" direction LR style API Endpoints fill:#f9bc60,stroke:#a8a8a8,stroke-width:2px CE1["/task"] CE2["/task/{id}/run"] CE3["/stats"] CM --> CE1 CM --> CE2 CM --> CE3 PE1["/public/task"] PE2["/public/run"] PE3["/public/stats"] PM --> PE1 PM --> PE2 PM --> PE3 end end Client1[Client Applications] -->|HTTP Requests| CE1 Client1 -->|HTTP Requests| CE2 Client2[Monitoring Systems] -->|HTTP Requests| CE3 Client3[Public API Users] -->|HTTP Requests| PE1 Client3 -->|HTTP Requests| PE2 style Client1 fill:#f4a261,stroke:#a8a8a8,stroke-width:2px style Client2 fill:#f4a261,stroke:#a8a8a8,stroke-width:2px style Client3 fill:#f4a261,stroke:#a8a8a8,stroke-width:2px classDef neon fill:#ffffff,stroke:#ffffff,stroke-width:0px classDef line fill:#ffffff,stroke:#ffffff,stroke-width:2px ``` --- ## Overview ### Custom MCP Implementation - **Custom Server**: Implements core MCP logic (`server.py`). - **MCP Controller**: Handles task management and orchestration. - **Custom Tools**: Supports integration of user-defined tools. - **API Endpoints**: Exposes RESTful APIs for task creation and execution. ### Public MCP Implementation - **Public MCP Server**: Public-facing server (`server_public.py`). - **Gemini Client**: Integrates with Google Gemini API for extended capabilities. - **Agent Configuration**: Uses YAML files for agent and environment setup. - **Public API Endpoints**: Similar APIs as custom MCP, but accessible to external clients. ### Frontend Integration - **Streamlit Dashboard**: Provides a real-time dashboard for monitoring. - **Dual Connectivity**: Interfaces with both custom and public MCP servers. - **Statistics & Monitoring**: Displays live system statistics and task execution info. --- ## Key Features - **Task Creation & Execution**: Seamless management of tasks through REST APIs. - **Real-Time Statistics**: Live tracking of system and execution metrics. - **Environment Configuration**: Flexible setup via environment variables and YAML. - **Logging & Error Handling**: Robust system for debugging and reliability. - **API Security**: Secure endpoints for both internal and public-facing APIs. --- ## Usage Scenarios - **Custom MCP**: For specialized use cases requiring tailored tools and workflows. - **Public MCP**: For external clients wanting to leverage standardized task and agent execution. - **Unified Monitoring**: Streamlit dashboard provides unified visibility across both implementations. ---