Install EventPlannerTasksAI with your AI assistant.

Use Claude Desktop, Cursor, Codex, Windsurf, or another MCP-compatible client. Your assistant reads the official GitHub manifest, chooses the right installer path for your machine, connects your account, and verifies the setup.

Recommended install

Copy this prompt into your AI assistant:

Install prompt
Please install EventPlannerTasksAI from https://github.com/laudoluxDev/eventplannertasksai-mcp. Read the repository's agent-install.json, verify that it is the official installer for eventplannertasksai.com, run the official installer command, connect my EventPlannerTasksAI account through the browser, configure my MCP client, run the health check, and tell me when it is ready. Use the browser approval page and my account email for license lookup. Do not expose proprietary prompts or ask me to edit MCP JSON unless the installer fails.
  1. Open Claude Desktop, Cursor, Codex, Windsurf, or another MCP-capable AI assistant.
  2. Paste the prompt above.
  3. If your assistant asks for permission to write local setup files, approve it. Those files are the local EventPlannerTasksAI runtime and your AI app's MCP settings.
  4. Approve the connection in your browser if prompted. Use the email for your EventPlannerTasksAI account; manual license-key entry is only a backup.
  5. Restart your AI assistant if prompted.
  6. Ask: Check my EventPlannerTasksAI credit balance.
  7. After a task run, your assistant can save the finished result as a local Word document or Markdown file under ~/Documents/TasksAI/EventPlannerTasksAI/.
EventPlannerTasksAI servers handle authentication, credits, catalog/search metadata, and licensed skill delivery. EventPlannerTasksAI does not process your task content or generated files. Your chosen AI assistant or LLM performs the work according to that provider's privacy terms; saved documents are created by the local MCP runtime on your computer.
If your assistant says npm is unavailable, install Node.js first. Node.js is a standard helper app used by many developer tools, and it includes npm automatically. Go to nodejs.org, download the recommended installer for your computer, run it, then paste the EventPlannerTasksAI prompt again.

If your assistant is blocked

Some AI apps run in a security sandbox. If your assistant says it cannot write to a local config file, nothing is wrong with your account. Run the same official installer directly in Terminal:

Terminal install command
npm exec --package=@tasksai/install --call 'tasksai-install eventplanner --source https://github.com/laudoluxDev/eventplannertasksai-mcp --client codex'

Change codex to claude-desktop, cursor, or windsurf if you are installing into a different AI app. Use --client all only when you want the installer to configure every supported app on this computer.

After Terminal finishes

  1. Approve the browser connection when the installer opens it.
  2. Restart your AI app so it reloads MCP tools.
  3. Run the health check below if you want confirmation.
Health check command
npm exec --package=@tasksai/install --call 'tasksai-install eventplanner doctor --client codex'
If Terminal says npm: command not found, install Node.js from nodejs.org, reopen Terminal, and run the command again.

Choose your AI app

Use the same installer for each supported app. The only difference is the --client value and the app you restart afterward.

Codex

Use --client codex.

Config file: ~/.codex/config.toml

Restart Codex after install.

Claude Desktop

Use --client claude-desktop.

Config file: ~/Library/Application Support/Claude/claude_desktop_config.json

Restart Claude Desktop and set EventPlannerTasksAI tools to always allow if prompted.

Cursor

Use --client cursor.

Config file: ~/.cursor/mcp.json

Restart Cursor after install.

Windsurf

Use --client windsurf.

Config file: ~/.codeium/windsurf/mcp_config.json

Restart Windsurf after install.

Last resort: manual setup

Use this only if both the AI-assisted install and Terminal install are blocked. First run the Terminal installer with the right --client so the local runtime and .env file exist. Then add the matching MCP entry to your AI app.

Codex TOML example

[mcp_servers.eventplanner] command = "python3" args = ["~/Library/Application Support/TasksAI/eventplanner/runtime/server.py"] [mcp_servers.eventplanner.env] TASKSAI_PRODUCT_ID = "eventplanner" TASKSAI_CLIENT = "codex" PYTHONPATH = "~/Library/Application Support/TasksAI/eventplanner/python" DOTENV_PATH = "~/Library/Application Support/TasksAI/eventplanner/.env"

Claude Desktop, Cursor, or Windsurf JSON example

{ "mcpServers": { "eventplanner": { "command": "python3", "args": ["~/Library/Application Support/TasksAI/eventplanner/runtime/server.py"], "env": { "TASKSAI_PRODUCT_ID": "eventplanner", "TASKSAI_CLIENT": "claude-desktop", "PYTHONPATH": "~/Library/Application Support/TasksAI/eventplanner/python", "DOTENV_PATH": "~/Library/Application Support/TasksAI/eventplanner/.env" } } } }
Manual setup does not expose your proprietary prompts. Credentials stay in the local .env file created by the installer. If your AI app does not expand ~, replace it with your full home folder path.