MCP Server

Plug Architecture Into Your AI Coding Tool

The Agents But Why MCP server gives your AI assistant direct access to our full knowledge base — 7 architecture patterns, 4 real-world use cases, and detailed component breakdowns. Ask questions, get recommendations, and make architecture decisions without leaving your editor.

⚡

Quick Start

The server is live — no installation required. Just add the endpoint to your AI coding tool:

https://www.agentsbutwhy.com/api/mcp

What is MCP?

The Model Context Protocol is an open standard that connects AI assistants to external tools and data sources. Instead of copy-pasting architecture docs into your chat, MCP lets the AI read from them directly — searchable, structured, always up to date.

🔍

5 Tools

Search, detail, recommend

📚

4 Resources

Browse the full catalog

💬

2 Prompts

Review & compare workflows

Available Tools

Tools are actions your AI assistant can invoke on your behalf.

search_patterns

Search architecture patterns by keyword, tag, or complexity level.

Try: "Show me patterns related to security"

get_pattern_detail

Get the full breakdown of a pattern — components, trade-offs, when-to-use.

Try: "Tell me everything about agentic-rag"

search_use_cases

Search real-world use cases by industry, complexity, or pattern.

Try: "Show me finance use cases"

get_use_case_detail

Get the full analysis of a specific use case.

Try: "Walk me through the AI Call Center use case"

recommend_architecture

Describe your project and get architecture pattern recommendations.

Try: "I'm building a customer support chatbot that needs to query our knowledge base..."

Browsable Resources

Resources are data your AI assistant can read for context — the full knowledge base exposed via URI.

agentsbutwhy://patternsComplete catalog of all architecture patterns

agentsbutwhy://patterns/{slug}Full detail of a specific pattern

agentsbutwhy://use-casesComplete catalog of all use cases

agentsbutwhy://use-cases/{slug}Full analysis of a specific use case

Connect Your Tool

Add the MCP server to your AI coding tool's configuration.

⌨️

Cursor

.cursor/mcp.json

{
  "mcpServers": {
    "agents-but-why": {
      "url": "https://www.agentsbutwhy.com/api/mcp"
    }
  }
}

🤖

Claude Desktop

claude_desktop_config.json

{
  "mcpServers": {
    "agents-but-why": {
      "url": "https://www.agentsbutwhy.com/api/mcp"
    }
  }
}

🚀

Antigravity

mcp_config.json

{
  "mcpServers": {
    "agents-but-why": {
      "serverURL": "https://www.agentsbutwhy.com/api/mcp"
    }
  }
}

🔌

Any MCP Client

MCP settings

Server URL: https://www.agentsbutwhy.com/api/mcp
Transport:  Streamable HTTP

Prompt Templates

Pre-built workflows for common architecture consultation scenarios.

🏗️

architecture_review

Describe your project and get a structured architecture review with pattern recommendations, phased adoption order, and trade-off analysis.

⚖️

pattern_comparison

Compare two or more architecture patterns side by side — strengths, weaknesses, synergies, and when to choose one over another.

How It Works

01

Add the endpoint

Copy the MCP server URL into your AI tool's settings. That's it — the server is already live.

02

Your AI connects

Your AI coding assistant discovers the available tools, resources, and prompts automatically via the MCP protocol.

03

Ask about architecture

Your AI assistant now has direct access to 7 architecture patterns, 4 use cases, and detailed component breakdowns. Just ask.

The MCP server is live at agentsbutwhy.com/api/mcp — connect from anywhere, no local setup required.