TasksBoard MCP server
Connect Claude, Cursor, VS Code, Codex or any MCP client to your real Google Tasks. Your agent reads, creates and organises tasks, and every change syncs to Google Tasks, your TasksBoard board and your phone.
Connector URL
https://mcp.tasksboard.com/mcp Free to connect · Sign in with Google · No API key, nothing to install
Setup in a minute
Pick your client, copy the setup and sign in with Google. The server is the same everywhere: one remote URL with Google sign-in.
In Claude, open Customize, then Connectors.
Click “+ Add”, then “Add custom connector”.
Name it TasksBoard, paste the connector URL and click Continue.
Keep the detected sign-in settings, click Add and sign in with your Google account.
https://mcp.tasksboard.com/mcp Works in Claude on the web, in Claude Desktop and in the mobile apps. The Free plan allows one custom connector. On Team and Enterprise plans an owner adds it under Organization settings, then members click Connect.
Run this command in your terminal. The user scope makes the server available in every project.
Sign in: run the login command, or type /mcp inside Claude Code and choose tasksboard. Your browser opens for the Google sign-in.
Ask Claude Code about your tasks.
claude mcp add --transport http tasksboard --scope user https://mcp.tasksboard.com/mcp claude mcp login tasksboard TasksBoard is already in the ChatGPT app directory, so there is nothing to configure. Add the app and sign in with Google.
TasksBoard for ChatGPTClick “Add to Cursor” and confirm the install. Or paste this into ~/.cursor/mcp.json (use .cursor/mcp.json inside a project to add it for that project only).
When Cursor connects to the server, complete the Google sign-in in your browser.
Ask the Cursor agent about your tasks.
{
"mcpServers": {
"tasksboard": {
"url": "https://mcp.tasksboard.com/mcp"
}
}
} Open the Command Palette and run MCP: Add Server.
Choose HTTP, paste the connector URL and name the server tasksboard.
For the destination choose Global, then Copilot Global. Or add the server to ~/.copilot/mcp-config.json yourself.
When VS Code asks, complete the sign-in with your Google account, then ask Copilot Chat about your tasks.
https://mcp.tasksboard.com/mcp {
"mcpServers": {
"tasksboard": {
"type": "http",
"url": "https://mcp.tasksboard.com/mcp",
"tools": [
"*"
]
}
}
} Run this command in your terminal.
Run the login command. Your browser opens for the Google sign-in.
Ask Codex about your tasks.
codex mcp add tasksboard --url https://mcp.tasksboard.com/mcp codex mcp login tasksboard The Codex CLI, the Codex IDE extension and the ChatGPT desktop app share this configuration, stored in ~/.codex/config.toml.
Add this to ~/.gemini/config/mcp_config.json (use .agents/mcp_config.json inside a workspace to add it for that workspace only).
In the Antigravity app, open Customizations, then Installed, and click Authenticate next to the server under MCP Servers. In the Antigravity CLI, type /mcp to open the MCP manager.
Sign in with your Google account, paste the authorization code if Antigravity asks for it, then ask the agent about your tasks.
{
"mcpServers": {
"tasksboard": {
"serverUrl": "https://mcp.tasksboard.com/mcp"
}
}
} Run this command in your terminal. The user scope makes the server available in every project.
Start Gemini CLI and run /mcp auth tasksboard. Your browser opens for the Google sign-in.
Ask Gemini about your tasks.
gemini mcp add --transport http --scope user tasksboard https://mcp.tasksboard.com/mcp /mcp auth tasksboard Since June 18, 2026 Gemini CLI serves only paid API keys and Gemini Code Assist Standard or Enterprise licences; with any other account, use Antigravity. The sign-in needs a browser on the same machine, so it does not work over a plain SSH session.
Any MCP client that supports remote servers works. Add a new server with the connector URL and these settings.
https://mcp.tasksboard.com/mcp Let the agent do it
Coding agents can add the server themselves. Paste one of these prompts, then sign in with Google when the browser opens.
Add the remote MCP server named tasksboard at https://mcp.tasksboard.com/mcp to your MCP configuration. It uses the Streamable HTTP transport and OAuth sign-in, so it needs no API key. When it is added, tell me how to sign in, then list my Google Tasks lists to confirm it works.
Run this command to add the TasksBoard MCP server: claude mcp add --transport http tasksboard --scope user https://mcp.tasksboard.com/mcp Then tell me how to sign in with claude mcp login tasksboard and whether I need to restart you before the tools load.
Add this server to ~/.cursor/mcp.json and keep the servers that are already there: {"mcpServers":{"tasksboard":{"url":"https://mcp.tasksboard.com/mcp"}}} Then tell me how to enable it and sign in with Google.
Add this server to ~/.copilot/mcp-config.json and keep the servers that are already there: {"mcpServers":{"tasksboard":{"type":"http","url":"https://mcp.tasksboard.com/mcp","tools":["*"]}}} Then tell me how to start it in VS Code and sign in with Google.
Run this command to add the TasksBoard MCP server: codex mcp add tasksboard --url https://mcp.tasksboard.com/mcp Then run codex mcp login tasksboard and tell me when the browser opens so I can sign in with Google.
Add this server to ~/.gemini/config/mcp_config.json and keep the servers that are already there: {"mcpServers":{"tasksboard":{"serverUrl":"https://mcp.tasksboard.com/mcp"}}} Then tell me how to authenticate it in Antigravity with my Google account.
Run this command to add the TasksBoard MCP server: gemini mcp add --transport http --scope user tasksboard https://mcp.tasksboard.com/mcp Then tell me to run /mcp auth tasksboard so I can sign in with Google.
Some agents load a new server only after a restart. The Google sign-in always happens in your browser: no agent can do that step for you.
After connecting
Ask in your own words. The agent picks the right tools and works in your live Google Tasks.
What's still open on my default list?
Show all my task lists and how many open tasks each one has.
Add “Send the Q3 report” to my Work list, due Friday, with a note to attach the spreadsheet.
Read the TODO comments in this repository and create a task for each one in a new list called “Tech debt”.
Break “Launch the new pricing page” into five subtasks.
Move everything due today that I haven't finished to tomorrow.
Mark the invoice task as done, then clear the completed tasks from that list.
Move the “Book flights” task from Personal to my “Trip to Lisbon” list.
12 tools
Everything you would do by hand in Google Tasks, available to your agent as MCP tools.
list_task_lists get_task_list create_task_list update_task_list delete_task_list list_tasks get_task create_task update_task delete_task move_task clear_completed_tasks Google Tasks stores due dates without a time of day, so a task created through the server has a due date but no due time.
Real Google Tasks
The server writes to your actual Google Tasks account, so whatever your agent does shows up everywhere you already work.
Your lists appear side by side as a kanban board on the web, with colours, drag and drop, and sharing for your team.
Tasks your agent creates show up in the Google Tasks side panel of Gmail, Calendar and Docs, and dated tasks land on your calendar.
The same tasks are in the Google Tasks app on Android and iOS, and in the TasksBoard mobile app.
Privacy
TasksBoard is a Google Cloud verified partner, ISO 27001 certified and GDPR compliant.
You authorise the connection on Google's own screen. Your password is never shared with the AI client or with TasksBoard, and there is no key to paste into a config file.
The server asks Google for your tasks and your basic profile (name and email). It has no access to your email, files or calendar events.
Remove the server from your AI client or revoke access in your Google Account settings, and the connection stops.
Everything about connecting AI agents to Google Tasks with TasksBoard MCP.
It is a remote Model Context Protocol (MCP) server for Google Tasks. AI agents such as Claude, Cursor, VS Code, Codex, Antigravity and Gemini CLI connect to it at https://mcp.tasksboard.com/mcp and can then read and manage your Google Tasks lists and tasks.
Yes. Connecting an AI agent and managing your tasks through it is free and does not need TasksBoard Premium. Premium is only about the extra features on the web board, such as list sharing, colours and backgrounds.
No. The server is hosted, so there is nothing to run on your machine. You add one URL to your AI client and sign in with Google in the browser.
Any MCP client that supports remote servers over Streamable HTTP with OAuth. This page has setup instructions for Claude, Claude Code, Cursor, VS Code, Codex, Antigravity and Gemini CLI. For ChatGPT, use the TasksBoard app in the ChatGPT app directory.
You do not need a separate account. You sign in with the Google account that holds your Google Tasks, and the server works straight away. The same account opens your board on tasksboard.com whenever you want to see your lists as a kanban board.
In your own Google Tasks account. The server reads and writes them through Google's official Tasks API; it does not keep a second copy of your task content.
It can set the day a task is due, but not the hour. The Google Tasks API stores a due date without a time of day.
Remove the server from the MCP settings of your AI client, or revoke access in the third-party apps section of your Google Account. Your tasks stay exactly where they are.
Add the TasksBoard MCP server once and keep your Google Tasks up to date from any AI client.