Connecting timr to an AI assistant (MCP Connector)

Overview

With the timr connector, you connect your timr data to an AI assistant and ask your questions about working time, project hours, and time account directly in the chat. The connector is a so-called MCP server that compatible AI assistants like Claude can integrate.

The assistant uses it to retrieve your time entries, project times, time accounts, and reports, and can also add entries on request, such as a note on a time entry.

You set up the connection once per assistant, after that, it remains permanently available.
 

Note: Data protection and security

The MCP connector accesses data in timr strictly within the permissions of the respective timr user. Before using it, clarify which data protection and security requirements apply to the use of AI assistants in accordance with your company's guidelines.

1. Prerequisites

2. Connecting timr to the AI assistant

2.1 In Claude
2.2 Integrating the timr MCP server directly

3. Asking questions and adding entries

3.1 Requesting reports
3.2 Adding notes and entries

1. Prerequisites

• An active timr account with login credentials.
• An AI client that supports MCP (e.g. Claude).
• The connector can also be used by users without additional permissions. A separate approval by an administrator is not required.

Note: Which reports the assistant can provide depends on your timr plan and the activated add-ons. Certain reports, such as those for overtime or subsidies, are only available if the corresponding add-on is active in your account.

2. Connecting timr to the AI assistant

Setup takes place in the respective AI assistant and is done once per assistant. The exact names of the menu items may differ depending on the version of the assistant.
 

2.1 In Claude

Step 1: In Claude, open Settings → Connectors.

Step 2: Select the timr connector.

Step 3: Click Connect and log in with your timr credentials.

2.2 Integrating the timr MCP server directly

timr is also available as an MCP server at the following address:

https://mcp.timr.app/mcp

If you are familiar with MCP, you can integrate this endpoint directly into an MCP-capable client of your choice — for example via the terminal or command line. Setup works like any other MCP server. You can find a brief introduction in the official MCP documentation.

3. Asking questions and adding entries

After connecting, you ask your questions in natural language. The connector is an AI-powered feature. The quality of the answers depends on how precisely you ask and how cleanly your data is documented in timr — for example, through a well-thought-out task structure, clear project names, and consistently recorded times. The more precise the question, the more accurate the answer.
 

3.1 Requesting reports

• "How many hours have I already worked this week, and how much is still outstanding?"
• "How many hours did we book on the Sample Customer project this month — broken down by task?"
• "What is my flextime balance, and how many vacation days do I have left this year?"
   

3.2 Adding notes and entries

You can also dictate entries to the assistant. Describe as precisely as possible which booking the entry belongs to.

"For my last time entry for customer Huber, task Service, add the note: Serviced heating system, replaced filter, customer was on site."

Tip: If you set up additional connectors — such as calendar or email — the assistant can combine information from multiple sources, for example to find free time slots taking into account your absences recorded in timr. Actions such as sending an email are only carried out by the assistant after your confirmation.