home / mcp / multi-agent tools platform mcp server
Provides five core tools and three agents with a FastAPI API and supervisor toolbox for coordinated tasks.
Configuration
View docs{
"mcpServers": {
"kingrishabdugar-mcp_demo": {
"command": "python",
"args": [
"tools_server.py"
]
}
}
}You are setting up a modular MCP server that coordinates five core tools and three specialized agents to perform advanced math, research, weather, and summarization tasks. It exposes a FastAPI-based API for interaction and includes a smart supervisor toolbox to orchestrate tools and agents. This guide walks you through how to install, start, and use the server locally with an MCP client.
Start all components in separate terminals to run the full MCP stack. The server provides an API surface for each agent and tool, allowing you to perform complex workflows by composing math, search, weather, and summarization capabilities with the specialized agents for research, math, and meteorology. Use an MCP client to interact with the supervisor tool exposed by the toolbox, or invoke individual agents/tools through their corresponding endpoints. You can mix tool invocations and agent reasoning to tackle tasks such as multi-step research, data analysis, weather-informed planning, and automated summarization.
Prerequisites: you need Python installed and a shell to create and activate a virtual environment.
cd C:/Users/kingr/CascadeProjects
python -m venv venv
venv\Scripts\activate
pip install -r requirements.txtLaunch each component in its own terminal to run concurrently.
python tools_server.py
python api_servers.py
python agent_tools.pyThe FastAPI server listens on http://127.0.0.1:8000 and exposes endpoints for each tool and agent. Use an MCP client to interact with the supervisor tool in the agent tools component.
Environment variables are used for deployment and API configuration as needed by the runtime. If you plan to deploy to a cloud environment or integrate with external services, set variables such as deployment names or API versions as required by your setup.
After starting all components, verify that the API is reachable and that you can invoke tools and agents via the MCP client. Run simple workflows to ensure the Math, Search, Weather, Wikipedia, and Summarizer tools respond correctly, and that the Research, Math, and Meteo agents perform as expected when invoked through the supervisor.
All runtime commands are executed locally in separate terminals. If you modify or extend the setup, ensure that dependencies in requirements.txt remain compatible and that the virtual environment is activated before running any component.
Core mathematical computations and reasoning tool used by the agents to perform calculations, algebra, calculus, and numeric analysis.
Web search and information retrieval tool used to gather up-to-date data and sources for agent reasoning.
Weather data tool to fetch forecasts and meteorological information for planning and analysis.
Knowledge lookups from encyclopedic sources to provide factual context and summaries.
Summarization tool to condense long texts into concise outputs for reports and briefs.
Specialized agent designed to perform multi-step research tasks, aggregating sources and forming informed conclusions.
Agent focused on mathematical problem solving, proofs, and symbolic reasoning using the Math tool.
Agent tailored to weather-informed tasks, integrating weather data into decision workflows.
Unifies tool and agent endpoints, coordinating tasks through the MCP toolbox.