Skip to main content
The Take Time MCP Server is distributed as an npm package. You configure it by adding a JSON block to your AI client’s settings file — no separate process to manage. The server reads your API key from an environment variable and handles authentication automatically.

​
Prerequisites

Before you begin, make sure you have:
  • Node.js 18+ — required to run the server via npx. Check your version with node --version. Download Node.js from nodejs.org if needed.
  • A Take Time API key — generate one from Settings → Developer → Generate API Key inside the Take Time app. Your key starts with tt_live_.
You do not need to install the MCP Server package globally. The npx command in the configuration fetches and runs it on demand each time your AI client starts.

​
Setup by Client

Choose the client you want to connect and follow the steps below.

​
Claude Desktop

1

Open the Claude Desktop configuration file

In Claude Desktop, go to Settings → Developer → Edit Config. This opens the claude_desktop_config.json file in your default editor.
2

Add the Take Time server block

Inside the mcpServers object, add the following entry. If mcpServers does not exist yet, create it at the top level of the JSON file.
{
  "mcpServers": {
    "taketime": {
      "command": "npx",
      "args": ["@taketime/mcp-server"],
      "env": {
        "TAKETIME_API_KEY": "tt_live_your_key_here"
      }
    }
  }
}
Replace tt_live_your_key_here with your actual API key.
3

Save and restart Claude Desktop

Save the configuration file, then fully quit and relaunch Claude Desktop. Changes to MCP configuration only take effect after a restart.
4

Confirm the tools are loaded

Look for the hammer icon (🔨) in the Claude Desktop chat toolbar. Clicking it shows a list of connected tools — you should see the Take Time tools listed there. If the icon is missing, double-check your JSON syntax and API key.

​
Cursor

1

Open the Cursor MCP configuration file

Create or open ~/.cursor/mcp.json. You can do this directly from the terminal:
open ~/.cursor/mcp.json
2

Add the Take Time server block

Add the following to the file, merging with any existing mcpServers entries:
{
  "mcpServers": {
    "taketime": {
      "command": "npx",
      "args": ["@taketime/mcp-server"],
      "env": {
        "TAKETIME_API_KEY": "tt_live_your_key_here"
      }
    }
  }
}
Replace tt_live_your_key_here with your actual API key.
3

Restart Cursor

Fully quit and reopen Cursor. MCP servers are initialized at startup.
4

Verify the connection

Open the Cursor AI panel and ask: “What study blocks do I have today?” If Cursor can reach Take Time, it will respond with your actual schedule.

​
VS Code

1

Open the VS Code MCP configuration file

You can configure MCP servers either globally in your user settings.json or per-project in .vscode/mcp.json. For a project-scoped setup, create .vscode/mcp.json in your workspace root.
2

Add the Take Time server block

Add the following configuration:
{
  "mcpServers": {
    "taketime": {
      "command": "npx",
      "args": ["@taketime/mcp-server"],
      "env": {
        "TAKETIME_API_KEY": "tt_live_your_key_here"
      }
    }
  }
}
Replace tt_live_your_key_here with your actual API key.
3

Reload the VS Code window

Open the Command Palette (Cmd+Shift+P / Ctrl+Shift+P) and run Developer: Reload Window to apply the new configuration.
4

Verify in GitHub Copilot Chat

Open the Copilot Chat panel and ask: “What study blocks do I have today?” A successful response confirms the MCP Server is connected and authenticated.

​
Verify Your Setup

Regardless of which client you use, the fastest way to confirm everything is working is to ask your AI assistant:
“What study blocks do I have today?”
If the MCP Server is connected and your API key is valid, the assistant will call list_blocks with today’s date and return your actual schedule. If you see a tool-not-found error or an authentication failure, revisit the configuration steps above and confirm that your API key starts with tt_live_.
Do not commit your mcp.json or claude_desktop_config.json to a public repository if it contains your API key. Consider using a secrets manager or environment variable injection for team environments.