Architect

v1.0.0Coding Agentsstable

A powerful, self-extending MCP server for dynamic AI tool orchestration. Features sandboxed JS execution, capability-based security, automated rate limiting, marketplace integration, and a built-in monitoring dashboard. Built for the Model Context Pr

aiai-agentsaitoolsautomationclaude
Share:
25
Stars
0
Downloads
0
Weekly
0/5

What is Architect?

Architect is a Model Context Protocol (MCP) server that allows AI assistants like Claude, Cursor, and VS Code to powerful, self-extending mcp server for dynamic ai tool orchestration. features sandboxed js execution, capability-based security, automated rate limiting, marketplace integration, and a built-in moni...

A powerful, self-extending MCP server for dynamic AI tool orchestration. Features sandboxed JS execution, capability-based security, automated rate limiting, marketplace integration, and a built-in monitoring dashboard. Built for the Model Context Pr

This server falls under the Coding Agents category on MCPgee, the world's largest MCP server directory with 33,000+ servers.

Features

  • A powerful, self-extending MCP server for dynamic AI tool or

Use Cases

Dynamic tool orchestration
Sandboxed code execution
Rate limiting and monitoring
ageborn-dev

Maintainer

LicenseMIT
Languagetypescript
Versionv1.0.0
UpdatedMar 22, 2026
Statushealthy
Maintenanceactive

Works with

ClaudeOpenAIwindowsmacoslinux

Installation

Manual Installation

npx architect

Configuration

Configuration Details

Config File

claude_desktop_config.json

Performance

Response Metrics

Response Time< 200ms
ThroughputMedium

Resource Usage

Memory UsageLow
CPU UsageLow

How to Set Up and Use Architect

Architect MCP Server is a self-extending AI tool orchestration platform that lets Claude create, store, and execute custom JavaScript tools at runtime inside a sandboxed environment. It provides 8 gateway tools covering tool creation and testing, batch execution, cron scheduling, webhook automation, secret storage, marketplace sharing, and a browser automation gateway — all managed through a real-time web dashboard. Teams building AI agent workflows use it to extend Claude's capabilities dynamically without redeploying their MCP server.

Prerequisites

  • Node.js installed (for local setup) OR Docker and Docker Compose (for containerized setup)
  • An MCP-compatible client such as Claude Desktop, Claude Code, or Cursor
  • Optional: GitHub personal access token for marketplace tool sharing features
1

Install via npm or Docker

Choose your preferred deployment method. The npm approach runs the server directly with Node.js; the Docker approach isolates it in a container with persistent volumes.

# Option A: npm (local)
npm install
npm run build
npm start

# Option B: Docker
docker compose up -d
2

Configure environment variables (optional)

Tune sandbox behavior by setting environment variables before starting the server. These control child process pool size, memory limits per sandbox, and audit log retention.

# .env or shell environment
ARCHITECT_SANDBOX_POOL=4
ARCHITECT_SANDBOX_MAX_EXEC=50
ARCHITECT_SANDBOX_MEMORY_MB=128
ARCHITECT_AUDIT_MAX_ROWS=10000
3

Add the server to your MCP client configuration

Register Architect in your Claude Desktop, Claude Code, or Cursor config file using npx to run the server package.

{
  "mcpServers": {
    "architect": {
      "command": "npx",
      "args": ["architect-mcp-server"]
    }
  }
}
4

Open the monitoring dashboard

After starting the server, navigate to the built-in web dashboard at http://localhost:3001 to monitor tool executions, view audit logs, and manage your tool library in real time.

5

Create your first custom tool

Ask Claude to create a new tool using the tool gateway. Provide a name, description, and JavaScript implementation. The tool is validated, sandboxed, and made immediately available for subsequent calls.

Architect Examples

Client configuration

Claude Desktop configuration for the Architect MCP Server using npx. Set environment variables for sandbox tuning if needed.

{
  "mcpServers": {
    "architect": {
      "command": "npx",
      "args": ["architect-mcp-server"],
      "env": {
        "ARCHITECT_SANDBOX_POOL": "4",
        "ARCHITECT_SANDBOX_MEMORY_MB": "128"
      }
    }
  }
}

Prompts to try

Sample prompts that use Architect's self-extending capabilities to create tools, run batches, and set up automation.

- "Create a new tool called 'fetch-weather' that calls the Open-Meteo API and returns the current temperature for a given city"
- "List all the custom tools currently available in Architect"
- "Run the 'fetch-weather' tool for these 10 cities in a batch: New York, London, Tokyo, Berlin, Sydney"
- "Schedule the 'daily-digest' tool to run every day at 8 AM using a cron expression"
- "Store my GitHub token as a secret called GITHUB_TOKEN so tools can access it securely"

Troubleshooting Architect

Sandbox execution errors or tools failing with memory exceeded

Increase the ARCHITECT_SANDBOX_MEMORY_MB environment variable from the default 128 MB. If multiple tools run concurrently, also increase ARCHITECT_SANDBOX_POOL to add more child processes. Set these in your shell environment or a .env file before starting the server.

Dashboard at localhost:3001 is not accessible

The dashboard starts automatically with the server on port 3001. If running in Docker, ensure port 3001 is mapped in your docker-compose.yml. If running locally, check that no other process is using port 3001 with lsof -i :3001.

Webhook server not available at port 3002

The webhook server on port 3002 starts on demand when a webhook automation is configured, not at server startup. Create a webhook trigger via the automate gateway or dashboard first, then test the endpoint.

Frequently Asked Questions about Architect

What is Architect?

Architect is a Model Context Protocol (MCP) server that powerful, self-extending mcp server for dynamic ai tool orchestration. features sandboxed js execution, capability-based security, automated rate limiting, marketplace integration, and a built-in monitoring dashboard. built for the model context pr It connects AI assistants to external tools and data sources through a standardized interface.

How do I install Architect?

Follow the installation instructions on the Architect GitHub repository. Clone the repo, install dependencies, and add the server config to your AI client.

Which AI clients work with Architect?

Architect works with all major MCP-compatible AI clients including Claude Desktop, Claude Code, Cursor, VS Code (GitHub Copilot), Windsurf, and Cline.

Is Architect free to use?

Yes, Architect is open source and available under the MIT license. You can use it freely in both personal and commercial projects.

Browse More Coding Agents MCP Servers

Explore all coding agents servers available in the MCPgee directory. Each server includes setup guides for Claude, Cursor, and VS Code.

Quick Config Preview

{ "mcpServers": { "architect": { "command": "npx", "args": ["-y", "architect"] } } }

Add this to your claude_desktop_config.json or .cursor/mcp.json

Read the full setup guide →

Ready to use Architect?

Browse our complete directory of 33,000+ MCP servers, read setup guides for your editor, and start building with the Model Context Protocol.

33,000+ ServersFree & Open SourceStep-by-Step Guides