AI Expert Workflow MCP server for AI agents

AI Expert Workflow MCP is a server implementing the Model Context Protocol (MCP) that enables consulting with AI experts in product management, UX design, and software architecture to plan and develop projects. It offers a natural, conversational experience that guides you through the product development process, resulting in comprehensive documentation and specifications.

Installation

Global Installation (Recommended)

Install AI Expert Workflow MCP globally from npm:

npm install -g ai-expert-workflow-mcp

Getting an API Key

Get an OpenRouter API key from OpenRouter.

Configuration

Set your API key as an environment variable:

export OPENROUTER_API_KEY=your_key_here

Using the MCP Server

Standalone PRD Generation

The easiest way to use the system is to generate a PRD in one step:

# Set your API key
export OPENROUTER_API_KEY=your_openrouter_key_here

# Generate a PRD with one command
npx ai-expert-workflow-generate "I want to build a recipe app that helps users find recipes based on ingredients they already have at home. Target users are home cooks who want to reduce food waste and save money."

This will:

Generate a comprehensive PRD based on your description
Save it as prd.md in your current directory
Also save it in Task Master compatible format (at scripts/prd.txt)

Interactive Conversation Workflow

For a more interactive experience, you can engage in a conversation with the AI experts:

Step 1: Plan with the AI Product Manager

User: Can you start the AI Expert Workflow for my new mobile recipe app?

AI: I'll start the AI Expert Workflow to help you plan your mobile recipe app. What specific aspect would you like to begin with?

User: Let's create a PRD with the AI Product Manager. I want to build a recipe app that helps users find recipes based on ingredients they already have at home.

AI: [Asks clarifying questions about target users, key features, monetization, etc.]

User: [Provides detailed responses about the app vision]

Step 2: Generate the PRD

User: Can you generate a complete PRD document based on our consultation?

AI: [Generates comprehensive PRD with MVP focus and lean startup validation plan]

AI: Document saved to PRD.md. You can now review it and use it for your development planning.

Integration with Cursor AI

To use the MCP server with Cursor AI, add this configuration to your Cursor settings:

{
  "mcpServers": {
    "ai-expert-workflow": {
      "command": "npx",
      "args": ["-y", "ai-expert-workflow-mcp"],
      "env": {
        "OPENROUTER_API_KEY": "YOUR_OPENROUTER_API_KEY_HERE",
        "OPENROUTER_MODEL": "tngtech/deepseek-r1t-chimera:free",
        "MAX_TOKENS": 8000,
        "TEMPERATURE": 0.7
      }
    }
  }
}

You can choose different AI models by changing the OPENROUTER_MODEL value:

For creative tasks: openai/gpt-4o, anthropic/claude-3-opus-20240229
For balanced performance: anthropic/claude-3-sonnet-20240229, mistral/mistral-large
For faster, cost-effective options: openai/gpt-3.5-turbo, anthropic/claude-3-haiku-20240307

See the complete list of OpenRouter models for all available options.

The Three-Phase Expert Workflow

The AI Expert Workflow implements a structured approach with three expert phases:

1. Product Definition (AI Product Manager)

The Product Manager phase covers:

Product Vision
User Personas
Business Requirements
Feature Map
Success Criteria

Output: A comprehensive Product Requirements Document (PRD)

2. UX Design (AI UX Designer)

The UX Designer phase covers:

UI Documentation
Feature Specifications
User Journeys
Interaction Patterns
Data Requirements

Output: A detailed UX Design Document

3. Technical Architecture (AI Software Architect)

The Software Architect phase covers:

Technical Architecture
API Specifications
Implementation Tasks
Database Schema
Testing Strategy

Output: A comprehensive Software Specification

Optional Task Master Integration

If you want to break down your PRD into development tasks, you can integrate with Task Master:

Install Task Master globally:
```
npm install -g task-master-ai
```
Get an Anthropic API key from Anthropic

Add both MCPs to your Cursor settings:

{
  "mcpServers": {
    "ai-expert-workflow": {
      "command": "npx",
      "args": ["-y", "ai-expert-workflow-mcp"],
      "env": {
        "OPENROUTER_API_KEY": "YOUR_OPENROUTER_API_KEY_HERE"
      }
    },
    "taskmaster-ai": {
      "command": "npx",
      "args": ["-y", "task-master-ai"],
      "env": {
        "ANTHROPIC_API_KEY": "YOUR_ANTHROPIC_API_KEY_HERE",
        "MODEL": "claude-3-sonnet-20240229"
      }
    }
  }
}

Parse your PRD with Task Master:

Can you parse the PRD at scripts/prd.txt and generate tasks?

How to add this MCP server to Cursor

There are two ways to add an MCP server to Cursor. The most common way is to add the server globally in the ~/.cursor/mcp.json file so that it is available in all of your projects.

If you only need the server in a single project, you can add it to the project instead by creating or adding it to the .cursor/mcp.json file.

Adding an MCP server to Cursor globally

To add a global MCP server go to Cursor Settings > MCP and click "Add new global MCP server".

When you click that button the ~/.cursor/mcp.json file will be opened and you can add your server like this:

{
    "mcpServers": {
        "cursor-rules-mcp": {
            "command": "npx",
            "args": [
                "-y",
                "cursor-rules-mcp"
            ]
        }
    }
}

Adding an MCP server to a project

To add an MCP server to a project you can create a new .cursor/mcp.json file or add it to the existing one. This will look exactly the same as the global MCP server example above.

How to use the MCP server

Once the server is installed, you might need to head back to Settings > MCP and click the refresh button.

The Cursor agent will then be able to see the available tools the added MCP server has available and will call them when it needs to.

You can also explictly ask the agent to use the tool by mentioning the tool name and describing what the function does.