Litak MCP: A Habit Tracker, Wheel of Life and Planner for Your AI Assistant

Connect your AI assistant to Litak through MCP. Read your Wheel of Life, check scheduled habits, manage daily tasks and log sleep with permissions you choose.

You tell your AI assistant you want a more balanced week. It suggests exercise, better sleep and time for a project. The advice sounds reasonable, but your actual goals, habit schedule and unfinished tasks are in another app. You have to explain them again, then copy the plan back.

Litak is a life planning app with a habit tracker, a 12-area Wheel of Life, goals, a weekly planner and sleep records. Its MCP server lets a compatible AI assistant read selected information and make permitted changes in your Litak account. You can discuss your week and update the plan in the same conversation.

For example, you can ask the assistant to read your saved Wheel of Life, show the habits scheduled for a particular date, mark a habit you have completed, set your main task for the day or record five hours of sleep. Each action depends on the access you grant and the tools your client supports.

What is MCP, and how does Litak use it?

MCP stands for Model Context Protocol. It is a standard for connecting AI applications to external data and tools, described in the official MCP introduction.

In this setup, Litak holds the records, the MCP server exposes a defined set of operations, and your AI client decides which operation to request from your conversation. The model can ask for the current weekly plan rather than work from an old copy pasted into a chat.

The connection address is https://litak.io/mcp, using Streamable HTTP. Litak is listed in the official MCP Registry as io.litak/mcp. The registry entry includes the server's connection address and authentication description.

You need a Litak account and an AI client that supports the required remote MCP connection and authentication method. The Litak MCP connection guide explains OAuth and dedicated Bearer keys.

What can an AI assistant do with Litak MCP?

The current server provides ten tools. Read permissions and write permissions are separate, including separate access to sleep records.

The assistant receives structured records through these tools. That does not mean it sees the dashboard as a screenshot or has unrestricted access to every feature in the app.

A habit tracker with MCP that follows your schedule

A useful AI habit tracker needs to distinguish a missed habit from a habit that was never planned for that day.

Suppose strength training is scheduled for Monday, Wednesday and Friday. On Thursday, asking the assistant to mark your completed habits should not create a Thursday training check. Litak exposes the scheduled dates alongside your habits, and the server skips attempts to add a completion outside that schedule. An incorrect old check can be undone separately.

Try a request with enough context to act on:

Show my habits planned for today. I completed my walk and reading; mark those two as done.

The assistant should first read the habits and the date, identify the matching records, then report the result of each change. If an action was skipped, it should say so. A response in the conversation is useful only when the corresponding record was actually updated.

To build a routine before connecting an assistant, see how to track habits. If you have a list of habits but no clear reason for them, start with connecting habits to goals.

Use your Wheel of Life as context for an AI weekly plan

The Wheel of Life is a self-assessment of different areas of your life. Litak uses 12 areas and saves your assessment so you can return to it. Through MCP, an assistant can read your latest saved result when you grant Wheel access.

That gives a planning conversation a specific starting point:

Read my latest Wheel of Life and this week's plan. Which area have I rated low but made no time for? Suggest one realistic action before adding anything.

The assistant can compare the information returned by the Wheel and planner tools. You then decide whether its interpretation fits your circumstances. A low score is your assessment of an area, not a diagnosis or proof of what caused the problem.

MCP currently provides read access to the saved Wheel result. Complete or change the assessment in Litak itself. For a practical route from a score to a weekly action, read Wheel of Life examples and how to plan your week.

Set the main task for today without creating a duplicate

A regular task dated today and the main task shown at the top of the Litak dashboard have different roles. An assistant needs to choose the right operation.

Make the existing “Prepare the launch” task my main task for today. Keep it as the same task in the weekly planner.

The intended sequence is to read the weekly plan, find the existing task and select it with set_task_of_day. If the task does not exist, the assistant can create it after you agree, then select it. Creating a regular task alone does not set the main daily task.

This matters when your assistant already holds a long conversation about a project. You can bring one existing action into focus without asking it to recreate your task list.

Log sleep as hours, separately from habit checkboxes

Sleep is a numeric record. “I slept five hours” carries different information from “I completed my habits.”

Record five hours of sleep for today, then show my sleep entries for this month.

With the sleep permissions enabled, set_sleep_entry records the hours and get_sleep reads the saved entries. The current input accepts whole hours from 0 to 24. Repeating a request for the same date updates that entry. A date with no entry means no value was recorded; it does not mean zero hours of sleep.

An assistant should not guess sleep hours from a bulk request to complete habits. You supply the number and the calendar date. When “today” could be ambiguous, include the date explicitly.

How to connect an AI assistant to Litak

1. Open the MCP setup guide and check that your AI client supports remote Streamable HTTP MCP and one of Litak's authentication methods.
2. Add https://litak.io/mcp as the server address.
3. Use OAuth where supported, or create a dedicated key under Settings → AI connections and configure the client with Bearer authentication. Keep the key in the client's credential settings.
4. Choose the permissions for your use case. A weekly review can start with read access; changing tasks, habits or sleep needs the corresponding write access.
5. Check the tools visible in your client. Start with a read request, such as showing this week's plan, before asking it to update a record.

If a newly added tool is missing, compare the client's tool list with the current server catalog. A stale client registration can keep showing an older list after a server update. Follow that client's refresh or reconnection procedure; do not ask the model to substitute a different write operation.

Who is this useful for?

Litak MCP fits a workflow where you already use an AI assistant to think through your week and want that conversation connected to your actual plan. The habit schedule, goals and saved Wheel provide context; permitted task and check-in operations let you record an agreed action without copying it between tools.

It is also useful for a weekly review with read access only. You can discuss what was planned, what was completed and which area deserves attention, then make changes yourself in Litak.

If you only want a printable tracker or a standalone assessment, the habit tracker PDF or Wheel of Life page may be enough. The MCP connection becomes useful when you want your assistant to work with the records you keep over time.

Start with one request: connect using the Litak MCP guide, ask for your current weekly plan, and check that the returned tasks match what you see in Litak. Then grant write access only for the actions you want to take through your assistant.