Connecting the MCP Server
What is an MCP Server
MCP (Model Context Protocol) is an open standard that lets an AI call the tools of external services directly. By connecting UTasks as an MCP server, you give your AI agent the ability to manage your tasks in real time, instead of just describing them manually.
Once the MCP server is connected, a number of practical scenarios become possible.
The AI agent can show you all tasks assigned to you and filter them by project, status, priority, or date range — for example, "show me tasks for the next 7 days" or "what's overdue." It can create new tasks and assign them right away, comment on tasks, mark them as done or reopen them, and pull up the list of project members and projects so you don't have to track names and IDs manually.
This is especially handy when UTasks is your main work tool: instead of switching between Telegram and your AI chat, you can handle your entire task workflow in one place, while the AI agent fetches up-to-date data from UTasks on its own.
How to Connect (We'll walk through it using Claude as an example)
Open your profile in the AI and go to Settings.
Then ConnectorsAddAdd custom connector.
In the connector settings window, enter a name for the server (whatever's convenient for you) and the address https://app.utasks.io/utb/mcp.
Click Add.
In the next window, the AI agent will ask for an API key to connect to UTasks. Open UTasks and run the/apikey new command. Copy the key.
How to Use It
Once connected, you can give the AI agent commands like: "show me my tasks for today," "what tasks are overdue," "create a task and assign it to Viktor," "mark this task as done."
Depending on what your AI supports, you could, for example, set up a daily morning summary of upcoming tasks, or an evening summary of what got done.
Good luck!