MCP.so
Sign In
A

Arkitect Mcp

@ezequielcasillas1

Repository

About Arkitect Mcp

Arkitect-MCP brings architecture-first reasoning to your AI coding workflow. Instead of jumping straight to refactors, the server scans repo context first — platform, workload, health, and intent — then returns actionable cursorGuidance for the agent to follow.

Basic information

Category

Reasoning

Transports

stdio

Publisher

ezequielcasillas1

Submitted by

ezequielcasillas1

Config

Add this server to your MCP-compatible client using the configuration below.

{
  "mcpServers": {
    "arkitect-mcp": {
      "command": "node",
      "args": [
        "packages/mcp-server/dist/stdio.js"
      ],
      "env": {
        "ARKITECT_DEFAULT_REPO_PATH": "C:\\Dev\\YourRepo"
      }
    }
  }
}

Tools

No tools detected

We auto-extract tools from the README. The maintainer can list them under a ## Tools heading to populate this section.

Overview

What is Arkitect MCP?

Arkitect MCP is an architecture-first reasoning server for AI coding workflows. It scans repository context—platform, workload, health, and intent—then returns actionable guidance for the AI agent. It integrates with Cursor via local stdio transport and is open source on GitHub.

How to use Arkitect MCP?

Clone the repository, build the server, and connect it to your MCP client (e.g., Cursor) using stdio transport. An optional Windows desktop app provides guided setup for Cursor. No specific configuration keys are mentioned in the README.

Key features of Arkitect MCP

  • Scans repo context before suggesting architecture changes
  • Provides repo diagnosis, architecture, pattern, and remix catalogs
  • Supports requirement tagging and refactoring analysis
  • Runs codebase verification and test suites
  • Exposes arkitect:// URIs for diagnosis, policy, and catalog data
  • Uses local stdio transport with Node.js

Use cases of Arkitect MCP

  • Adding a new feature to an existing codebase while preserving architectural consistency
  • Refactoring legacy systems with structured, context-aware guidance
  • Onboarding AI agents to repository patterns and design rules
  • Performing systematic codebase verification and automated test runs

FAQ from Arkitect MCP

What does Arkitect MCP do that a general AI coding assistant doesn’t?

Arkitect MCP injects architecture-first reasoning: it scans repository context before suggesting changes, rather than jumping straight to refactors. It returns structured cursorGuidance for the agent to follow.

How is Arkitect MCP installed and run?

It runs locally via Node.js with stdio transport. The README mentions cloning, building, and connecting to Cursor; an optional Windows desktop app simplifies Cursor integration.

What data does Arkitect MCP access?

It accesses repository context (platform, workload, health, intent) and provides arkitect:// URIs for diagnosis, policy, and catalog data. All processing is local.

Is Arkitect MCP open source?

Yes, it is open source on GitHub—clone, build, and connect freely.

What transport does Arkitect MCP use?

It uses local stdio transport via Node.js. The optional Windows desktop app provides a guided installation for Cursor.

Comments

More Reasoning MCP servers

MCP Sandbox

JohanLi233

Python sandboxes for llms

40

n8n Workflow Builder MCP Server

makafeli

AI assistant integration for n8n workflow automation through Model Context Protocol (MCP). Connect Claude Desktop, ChatGPT, and other AI assistants to n8n for natural language workflow management.

530

n8n - Secure Workflow Automation for Technical Teams

n8n-io

Fair-code workflow automation platform with native AI capabilities. Combine visual building with custom code, self-host or cloud, 400+ integrations.

195.5K

iFlytek Workflow MCP Server

iflytek

This a simple implementation of an MCP server using iFlytek. It enables calling iFlytek workflows through MCP tools.

28

Part 1. Real-Time LangGraph Agent with MCP Tool Execution

junfanz1

This project demonstrates a decoupled real-time agent architecture that connects LangGraph agents to remote tools served by custom MCP (Modular Command Protocol) servers. The architecture enables a flexible and scalable multi-agent system where each tool can be hosted independent

27