ai-software-architect
AI-powered architecture documentation framework with ADRs, reviews, and pragmatic mode. Now available as Claude Code Plugin for easiest installation.
claude mcp add --transport stdio codenamev-ai-software-architect mcp
How to use
The AI Software Architect MCP server provides a programmatic interface for integrating the AI Software Architect framework with other MCP-compatible tools and automation workflows. By running the MCP server, you can query and orchestrate architecture documentation, reviews, recalibration, and progress tracking through the MCP protocol. This enables automated project analysis, cross-tool collaboration, and streamlined execution of architecture workflows from external scripts or assistants such as Claude Code, Cursor, or other MCP-enabled clients. Use the server to plug in your own automation pipelines, create multi-project analyses, and leverage advanced project analysis features exposed by the MCP implementation.
To use it, run the MCP server via the configured command (mcp) and interact with it using MCP-compatible clients or the Claude/Cursor integrations outlined in your setup. The server acts as a programmatic endpoint for issuing architecture-related tasks, pulling context files, and returning structured results that can guide decision-making, documentation updates, and recalibration plans across your projects.
How to install
Prerequisites:
- Node.js installed (for npm-based installation) or a compatible runtime if using a different channel
- Access to the internet to fetch the npm package
Step-by-step installation:
- Install the MCP server package globally via npm:
npm install -g ai-software-architect
- Configure the MCP server in Claude settings (example path shown is ~/.claude/config.json):
{
"mcpServers": {
"ai-software-architect": {
"command": "mcp",
"args": []
}
}
}
- Verify installation by listing or invoking the server (example, depending on your environment):
mcp ai-software-architect --help
- Start using the MCP server with your MCP-compatible client or IDE integrations. For integration with Claude/Cursor, refer to your editor or plugin configuration as documented in the project guides.
Additional notes
Tips and common considerations:
- The MCP server name in your configuration is ai-software-architect and uses the mcp command with no extra arguments by default.
- Ensure Node.js (and npm) are up to date to minimize installation issues.
- If you encounter permission errors on macOS or Linux, consider using sudo for the npm install -g step or adjust npm prefix settings.
- When integrating with Claude or Cursor, you may need to configure credentials or authentication depending on your environment; consult the integration docs for any required environment variables.
- The MCP setup supports programmatic access; you can script tasks like architecture documentation generation, review orchestration, recalibration planning, and progress tracking across multiple projects.
- If you plan offline usage, check for offline-capable configurations or offline modes recommended in the project docs.
Related MCP Servers
buildwithclaude
A single hub to find Claude Skills, Agents, Commands, Hooks, Plugins, and Marketplace collections to extend Claude Code, Claude Desktop, Agent SDK and OpenClaw
skrills
Coordinate skills between Codex, Copilot, and Claude Code. Validates, analyzes, and syncs skills, subagents, commands, and configuration between multiple CLIs.
mcp
MCP server for EventCatalog
mcp-adr-analysis
A sophisticated Model Context Protocol (MCP) server for analyzing Architectural Decision Records (ADRs) and providing deep architectural insights to AI agents.
claude-image-gen
AI-powered image generation using Google Gemini, integrated with Claude Code via Skills or Claude.ai via MCP (Model Context Protocol).
clawfu-skills
172 expert marketing skills for AI agents — ClawFu MCP Server