Cut the tokens you send to an LLM, on your own machine.

1. Open Settings → Developer → Edit Config. That opens claude_desktop_config.json, and creates it if it isn't there yet. On a Mac it lives in ~/Library/Application Support/Claude/, on Windows in %APPDATA%\Claude\.
2. Add AsterMind under mcpServers. If the file already lists other servers, add the "astermind" entry beside them:

{
  "mcpServers": {
    "astermind": {
      "command": "npx",
      "args": ["-y", "@astermind/astermind-mcp"]
    }
  }
}

3. Quit Claude completely and open it again.
4. To check, click + at the bottom left of the chat box, then Connectors → Manage connectors. You'll see astermind with its 10 tools.

Run this in the project where you want it:

claude mcp add astermind -- npx -y @astermind/astermind-mcp

To use it in every project, add --scope user after add. Check with claude mcp list, which shows it as connected. The very first check can fail while npx downloads the package; run it again.

One command adds it:

codex mcp add astermind -- npx -y @astermind/astermind-mcp

Or add it to ~/.codex/config.toml yourself (on Windows, %USERPROFILE%\.codex\config.toml). The longer start-up timeout gives the first download time to finish:

[mcp_servers.astermind]
command = "npx"
args = ["-y", "@astermind/astermind-mcp"]
startup_timeout_sec = 30

Check with codex mcp list, or type /mcp in a Codex session. The Codex IDE extension and the ChatGPT desktop app, which runs Codex, read the same file.

Install it with pip install openai-agents and set OPENAI_API_KEY. The server runs on your machine, beside your agent:

import asyncio

from agents import Agent, Runner
from agents.mcp import MCPServerStdio


async def main() -> None:
    async with MCPServerStdio(
        name="astermind",
        params={
            "command": "npx",
            "args": ["-y", "@astermind/astermind-mcp"],
        },
        client_session_timeout_seconds=30,  # the default is 5 s; the first start downloads the package
    ) as server:
        agent = Agent(
            name="Assistant",
            instructions="Use the AsterMind tools when they help.",
            mcp_servers=[server],
        )
        result = await Runner.run(agent, "List the tools you can use.")
        print(result.final_output)


asyncio.run(main())

Install it with npm install @openai/agents zod, then:

import { Agent, run, MCPServerStdio } from "@openai/agents";

async function main() {
  const server = new MCPServerStdio({
    name: "astermind",
    command: "npx",
    args: ["-y", "@astermind/astermind-mcp"],
  });
  await server.connect();
  try {
    const agent = new Agent({
      name: "Assistant",
      instructions: "Use the AsterMind tools when they help.",
      mcpServers: [server],
    });
    const result = await run(agent, "List the tools you can use.");
    console.log(result.finalOutput);
  } finally {
    await server.close();
  }
}

main().catch(console.error);

Add this to ~/.cursor/mcp.json for every project, or to .cursor/mcp.json for one project, then restart Cursor. The server and its tools then appear in Cursor's MCP settings.

{
  "mcpServers": {
    "astermind": {
      "type": "stdio",
      "command": "npx",
      "args": ["-y", "@astermind/astermind-mcp"]
    }
  }
}

Add this to .vscode/mcp.json in your workspace, or run MCP: Open User Configuration from the Command Palette to use it everywhere. VS Code's top-level key is servers, not mcpServers.

{
  "servers": {
    "astermind": {
      "type": "stdio",
      "command": "npx",
      "args": ["-y", "@astermind/astermind-mcp"]
    }
  }
}

Or add it to VS Code in one click. Check with MCP: List Servers.

In LM Studio 0.3.17 or later, open the Program tab in the right sidebar and choose Install → Edit mcp.json. Add this block and save. LM Studio loads it straight away and asks before each tool call.

{
  "mcpServers": {
    "astermind": {
      "command": "npx",
      "args": ["-y", "@astermind/astermind-mcp"]
    }
  }
}

Or add it to LM Studio in one click.

Open Cline's MCP servers settings, choose to edit the configuration, and add the same block as for LM Studio above. In the Cline CLI, cline mcp walks you through it.

In Goose Desktop, open Extensions → Add custom extension, choose Standard IO, and enter the command npx -y @astermind/astermind-mcp. In the Goose CLI, run goose configure, then Add Extension → Command-line Extension with the same command.

Add a local (stdio) server with the command npx and the arguments -y @astermind/astermind-mcp.