MCP Server for Time Tracking: time cockpit in Claude, ChatGPT and Copilot

Ask Claude, ChatGPT, Copilot, Cursor or Codex about project budgets, open timesheets and utilization, in the assistant you already use.

Try for Free for 30 Days Setup in the documentation

Included in every plan · currently in public preview · requires Microsoft Entra ID

In the AI assistant

How is the Vitalmere budget, and what is still unbilled for September?

Patient portal
182 of 240 h (76%)
Unbilled
26.00 h
of which without a task
4.50 h

Shall I assign the task “Implementation” to the 4.50 h?

time cockpit MCP server

mcp.timecockpit.com

  1. get_current_user Sign-in and permissions checked
  2. execute_tcql_query Vitalmere projects and budgets
  3. get_timesheets September timesheets

The signed-in user's permissions apply.

Before every change

update_object Change 3 timesheets

Task
Implementation
Project
Patient portal
Duration
4.50 h

The time cockpit MCP server connects time tracking with AI assistants such as Claude or ChatGPT. MCP stands for Model Context Protocol: a standard that lets an assistant retrieve data from another application. The server exposes projects, clients, tasks, timesheets, absences and lists and runs at https://mcp.timecockpit.com.

Users sign in with their Microsoft Entra ID account, and their time cockpit permissions apply. You confirm creating, changing and deleting records beforehand. The MCP server is included in every plan and is currently in public preview.

  • Included Included in every time cockpit plan. License or usage costs apply only to your AI tool.
  • Quick to set up No server of your own required. Step-by-step guides for your IT cover the common AI tools.
  • In your control Your user permissions also apply to the assistant; changes are confirmed before they are executed.

Answers where you already work

Instead of building a report for every question, ask the question in your AI tool. The assistant fetches the data from time cockpit and summarizes it.

What the assistant can do with time cockpit

  • Automate workflows recurring checks and connections to other systems, without building an integration flow.
  • Analyze data questions about clients, invoicing and utilization in your own words, answered from your time cockpit data.
  • Steer projects budget consumption and status of your projects, without a dedicated report.
  • Track time prepare, review and create entries directly in your AI tool.

Your time cockpit permissions apply, including to custom fields and customizations.

Keep the AI tools you already use

The MCP server connects time cockpit with the common AI tools:

Claude
Claude Code, the Claude app, claude.ai and Cowork
OpenAI
Codex and ChatGPT in Business, Enterprise and Edu workspaces
Microsoft Copilot
Visual Studio Code with GitHub Copilot and Microsoft 365 Copilot through Copilot Studio
Cursor
editor and CLI

You do not need a new chat window or a separate AI vendor for time tracking. Ideally, use a tool your IT has already approved for company data. The documentation covers the setup for each tool.

The companion skills give the assistant guidance for time tracking, reports and data queries.

Public preview The MCP server is currently in public preview.

time cockpit MCP Server at a Glance

  • Included in every time cockpit plan; check possible AI costs separately
  • No server to run; setup and approval by your IT required
  • Sign-in with your own Microsoft work account; existing permissions apply
  • Read and write access possible; changes are confirmed before they are executed
Price
€10.85 per user per month
Trial
30 days free, no credit card
Hosting
Microsoft Azure, data centers in the EU
Single sign-on
Microsoft Entra ID

MCP server: queries, analyses and workflows in your AI assistant

The MCP server connects your AI assistants directly with time cockpit. Depending on the tool and your permissions, you can query data, create analyses and support your workflows.

Integration with your AI assistant

The MCP server makes time cockpit directly accessible to AI assistants. You work in Claude, ChatGPT, Copilot, Cursor or Codex and query information from your time tracking there. Each user signs in individually; existing permissions in time cockpit also apply to the assistant. Custom fields and customizations are accessible too, as far as your user permissions allow.

Query and analyze data

Your assistant can query projects, budgets, clients, tasks, timesheets and lists from time cockpit. It summarizes data, prepares report overviews and helps you interpret anomalies:

  • Project budgets: you see which projects have used a high share of their hour budget and where a conversation with the project manager is due.
  • Invoicing: you have billable timesheets without an invoice assignment compiled by client and project and bill them as usual through project invoicing.
  • Time tracking: you spot possible gaps and overlaps in timesheets before they complicate the month-end close.

Team leads ask about missing entries in the team, remaining vacation per employee and open vacation requests still awaiting approval.

Budget consumption of running projects

  • Webshop relaunch Meadowyn 115%
  • Patient portal Vitalmere 76%
  • Maintenance & support 2026 Wattbrook 70%
  • Site manager mobile app Brindlow 62%

The webshop relaunch is over budget and the patient portal has used three quarters. Discuss both projects with the project managers?

Prompt ideas for time tracking

  • “Create a budget traffic light for all open projects with consumption in percent and a traffic light status. Highlight changes compared to the previous week.”

    Weekly budget traffic light

    Set up as a scheduled task, the report is ready every Monday without anyone having to start an export.

  • “Create a dashboard for client Sample Ltd. covering the last twelve months: hours and revenue per month, distribution across projects, open invoices.”

    Client cockpit as an interactive dashboard

    In Claude this becomes an interactive dashboard with filters; in other assistants an HTML or Excel file. Ideal for preparing a client meeting.

  • “Here is the export of my Claude chats from last week. Derive possible work activities from it and compare them with my timesheets.”

    Derive timesheets from chat histories

    If you work with AI a lot, your chat history is a log of your work. The assistant suggests missing entries and flags uncertain assignments instead of inventing durations.

AI agents instead of integration flows: connect time cockpit with Jira and co.

Instead of building flows in n8n, Zapier or Power Automate, describe the workflow in your own words. Your AI agent carries it out with data from time cockpit and other connected systems, on a schedule if you like.

Workflows with AI agents

The MCP server is a technical foundation for workflows in which an AI assistant includes time cockpit data. Which steps an agent carries out on its own depends on the AI tool and your configuration. The agent prepares timesheets and other changes, and the hosted MCP server carries them out only after you confirm. Fully autonomous writes are therefore not intended.

Automate time tracking: workflows for your AI agent

  1. every Monday

    Budget alert before the project tips over

    Instead of finding out at month-end that a budget has been exceeded, you know while you can still talk to the client or replan.

    “Check all running projects every Monday. If a project has used more than 80 percent of its hour budget, send the project manager a short summary with the budget, booked hours and the tasks with the highest consumption.”

  2. every morning

    Report working time violations automatically

    Exceeded maximum daily working hours, missing breaks or rest periods that are too short stand out before they pile up. The basis is the standard list of working time violations in time cockpit.

    “Every morning, check yesterday's working time violations in my team and list them by employee and type of violation. Do not make any corrections.”

  3. whenever you like

    Book time directly from your AI agent

    You book where you are working, without opening time cockpit. The assistant shows the entry first and creates it only after you confirm. If you like, it checks for overlaps with existing entries beforehand.

    “Book 2.5 hours today on project Relaunch, task Concept, with the description Navigation workshop.”

  4. connected

    Connect tickets and time tracking without an integration flow

    If your assistant is also connected to Jira, Azure DevOps, Outlook or Teams, links between these systems come about without a flow of their own. For fixed interfaces with high data volumes, for example to ERP or payroll, the Web API remains the right choice.

AI time tracking: what the MCP server does and what it does not

Many vendors use “AI time tracking” for software that detects working time by itself and assigns it to projects. The MCP server does not do that. It gives the AI assistant of your choice access to time cockpit, within your permissions: the assistant answers questions, summarizes, proposes entries and creates them. The time itself still comes from your team.

If you would rather not reconstruct your day from memory, add the optional automatic time tracking with sidekick. It records on Windows and macOS what you worked on at your computer and helps you complete your own entries.

Connected in three steps

No server of your own, nothing in your own data center: time cockpit runs the MCP server. Your IT sets up sign-in and approvals, after that everyone connects with their own account.

  1. Your IT

    Create the app registration

    Your IT registers the connection in Microsoft Entra ID, Microsoft's user and access management. It sets up sign-in for the AI tools you want to use and grants the access permission. The technical steps are described in the Microsoft Entra ID guide.

  2. Your IT

    Connect your AI tool

    Your IT adds https://mcp.timecockpit.com and the sign-in configuration to the chosen tool. Depending on the tool, a central approval is also required. The guides below describe the steps for each tool.

  3. Each user

    Sign in and ask

    Each user signs in with their own Microsoft work account. Start with read access and a project question you know the answer to. Compare the answer with time cockpit.

The most important AI tools

Verify the connection

Security and control

The assistant works with your time cockpit permissions, and you decide how far it may go.

Personal sign-in
Everyone signs in with their own Microsoft account. The assistant sees exactly what you see in time cockpit.
Start read-only
With "readonly" you use the assistant for analysis only; "own data" limits it to your own records.
Confirm changes
When the assistant wants to create, change or delete records, it describes the change first and carries it out only after your approval.
Your tool, your choice
Your IT decides which AI tool is used, and with it the terms under which the data is processed.
Traceable results
Truncated results are marked, and every figure can be checked in time cockpit.

Technical details:

For your IT: setup and documentation

Sign-in, configuration per AI tool, access modes and limits are described step by step in the documentation.

Go to the documentation Set up sign-in

For your own integrations, the Web API remains available; for customizing the data model, there is the local MCP server in OnCockpit Admin. Use cases are covered in the post What AI Agents Can Do with Time Tracking; how we use AI in product development is described in Customizable Standard Software Meets AI.

Frequently Asked Questions

What is the time cockpit MCP server?

The MCP server is an interface based on the Model Context Protocol, the open standard for connecting AI assistants to applications. It runs at time cockpit at https://mcp.timecockpit.com. Your IT sets up sign-in and the connection to the AI tool. You do not have to run an MCP server of your own. Your AI tool connects to it and can read projects, clients, tasks, timesheets and lists from time cockpit and, if you want, create entries. time cockpit itself has no chat of its own: you work in the assistant you already use.

Which AI assistants are supported?

You can connect time cockpit with Claude Code, the Claude app, claude.ai, Cowork, OpenAI Codex, ChatGPT in company workspaces, Visual Studio Code with GitHub Copilot, Cursor and Microsoft 365 Copilot through Copilot Studio. The prerequisites and setup steps for each tool are in the documentation. The MCP server is currently in public preview.

What does the MCP server cost?

Nothing extra. The MCP server is included in every plan, including the free Freelancer year and the trial. The license for the AI tool itself, such as Claude or ChatGPT, is your agreement with that provider.

What are the prerequisites?

You need a Microsoft work account with access to time cockpit, an application registered by your IT in Microsoft Entra ID and a suitable AI tool. Microsoft Entra ID manages users and access. Your IT sets up sign-in and the connection. Step-by-step guides are available for the common AI tools, so your IT can get started right away.

Can the assistant change data in time cockpit?

Yes, within your user permissions. The documented default requires a confirmation when records are created, changed or deleted, including when time is booked. The assistant shows the planned change and passes the confirmation on to the server. This is not an independent approval process by a second person. If you want to rule out changes, connect the tool in “readonly” mode, that is with read access only.

Does time cockpit track time automatically with AI?

No. The MCP server does not detect working time by itself; it gives your AI assistant access to time cockpit. The assistant answers questions, summarizes data, proposes entries and creates them, while the time still comes from your team. To reconstruct your working day, the optional time cockpit sidekick records on Windows and macOS what you worked on at your computer.

Does the assistant see more than I do in time cockpit?

No. You sign in with your own Microsoft account, and every call is checked against your time cockpit permissions. What you cannot see in the web client, the assistant cannot see either. Access can additionally be limited to your own data, and internal system tables are never visible to assistants.

Where does my data go?

Everything the assistant reads through the MCP server is sent to the AI provider of the client and processed under that provider’s terms. Review your provider’s data retention, chat memory and training policies before you roll out access to your team. time cockpit logs every tool call and bounds result sizes.

Do I still need the Web API?

For your own integrations between applications, the Web API remains available. It is the programming interface of time cockpit. The MCP server makes data and functions accessible to AI assistants. The two complement each other; their feature sets are not identical. Model actions, for example, are not available through the hosted MCP server according to the documentation.

Does time cockpit fit your process, from time tracking to invoicing?

In 30 minutes you will know whether time cockpit covers your project hours, project controlling and invoicing.

Alexander HuberProduct Management and Implementation Michael KubitschkaProject Management and Implementation
What we cover
You tell us how you track and bill time today. We show you live how that works in time cockpit and what switching involves.
Who you talk to
Someone from the team that builds time cockpit and rolls it out for customers, for example Alexander or Michael.
How
30 minutes online via Microsoft Teams, free and without obligation.
Book a free 30-minute call or try it yourself, no credit card needed

You pick a free slot in Microsoft Bookings (opens in a new tab) and receive the confirmation with the Teams link by email. Prefer to write? support@timecockpit.com