ActiveCollab MCP Server

MCP (Model Context Protocol)

What is MCP and how does it work in ActiveCollab?

MCP (Model Context Protocol) allows AI assistants to securely connect to your ActiveCollab workspace and interact with your data.

When MCP is enabled, supported AI models can access information from your workspace and help you work faster. Depending on the permissions you allow, the AI can:

• Look up project and task details
  
• Answer questions about your projects
  
• Summarize discussions
  
• Find information across tasks and comments
  
• Create or update tasks (if write access is enabled)
  

For example, you could ask an AI assistant:

• Show me all overdue tasks in the Marketing project
• Which tasks are assigned to me this week?
• What's the status of the Website Redesign project?
• List all tasks without an assignee in the Q1 Campaign project
• How many tasks does each team member have this sprint?
• Who worked on the mobile app last week?
• Show me all tasks assigned to John that are due in the next 7 days
• How many hours did we log on the client project this month?
• Which tasks have no time tracked but are marked as complete?
• Show me a summary of time logged per team member in June
• Summarize the comments on the Homepage Redesign task
• Are there any unresolved discussions in the Sprint 12 project?
• What was the last update on the API integration task?

The Owner, both Member+ and Member roles of the ActiveCollab workspace can create and manage MCP tokens.

Once a token is created, it can be used to connect an AI client such as:

• Langdock

• Claude (Anthropic)

• Other MCP-compatible clients

After the connection is established, the AI can safely access the parts of ActiveCollab allowed by the token.

How to enable MCP in ActiveCollab?

1. Go to Apps & Integrations.
   
2. Open the MCP section.
   
3. Click Add MCP Client.
   
4. Enter your account password to confirm the action.
   
5. Choose the AI client you want to connect:
   
1. Langdock
2. Claude
3. Other
6. Configure the connection and generate the token.
7. Note: The Owner, both Member+ and Member roles of the ActiveCollab workspace can create and manage MCP tokens.

How to connect Langdock to ActiveCollab

If you are using Langdock, you can connect it to ActiveCollab by creating an MCP token and adding it as a remote MCP integration in Langdock.

▶

1. Go to Apps & Integrations → MCP.
   
2. Click Add MCP Client.
   
3. Enter your password.
   
4. Select Langdock.
   
5. (Optional) Enable Token is read-only if you want the AI to only view data.
   
6. Click Add.
   

ActiveCollab will generate:

• MCP Server URL
  
• MCP Token

How to connect Claude to ActiveCollab?

If you are using Claude Desktop, you can connect it to ActiveCollab using the MCP configuration snippet generated in ActiveCollab.

▶

1. Go to Apps & Integrations → MCP.
   
2. Click Add MCP Client.
   
3. Enter your password.
   
4. Select Claude.
   
5. Click Add.
   

ActiveCollab will generate:

• MCP Server URL
  
• MCP Token
  
• A configuration snippet for Claude Desktop

How to connect another MCP-compatible AI client?

If you want to connect another AI tool that supports MCP, you can generate a custom token and configure it in that client.

▶

If you are using another MCP-compatible client:

1. Go to Apps & Integrations → MCP.
   
2. Click Add MCP Client.
   
3. Enter your password.
   
4. Select Other.
   
5. Enter:
   
1. Client – the name of the application using the token
2. Vendor – the organization that built the application
6. Choose the access level:
   
1. Read Only – the AI can only view data
2. Read and Write – the AI can create and update items
7. Click Add.
   

ActiveCollab will generate:

Who can create MCP tokens?

Every Member and Owner of the ActiveCollab workspace can create MCP tokens.

Tokens are generated from the Apps & Integrations → MCP page.

If you need to have a clear distinction of what was performed by the agent, and what the regular users do, we recommend using a dedicated business account for the AI client instead of a personal account. This prevents sharing personal login credentials, and provides better audit trails.

Can multiple AI models use MCP at the same time?

Yes, you can use the same MCP server URL and token with multiple AI clients.

However, we recommend creating separate tokens for each client. This gives you better visibility and control over how each AI integration is used.

Creating separate tokens also makes it easier to:

• track usage
  
• manage permissions
  
• revoke access for a specific client if needed.
  

Use Source Code with AI agents

When GitHub or GitLab is connected to ActiveCollab, an AI agent can help you work with the source code linked to tasks. The Source tools follow the same permissions and client visibility rules as the app.

• Get a branch name for a task, instructions for referencing the task, and your mapped Git username with suggest_branch_name.
• Review the branches and pull requests linked to a task, including their pipeline and deploy state, with get_task_branches_and_pull_requests.
• Review the activity for one linked branch or pull request with get_branch_or_pull_request_history.
• Find the repositories connected to a project with project_repositories in the list tool.
• Link an existing branch or open pull request to a task with link_branch_or_pull_request_to_task. You need permission to edit the task.
• Find a person's mapped Git username for each repository with get_username_for_repositories.

The Git username comes from the user mapping that an administrator sets for a connection. See Map users.

Use Availability with AI agents

When Availability is turned on, an AI agent connected through MCP can tell you who is off and add time off for you or your team.

For example, you could ask:

• Who is off this week?
• How many days off did each person take this year?
• Add a vacation for Anna from July 7 to July 18.
• How many hours can Mark work next month?

The agent uses these tools:

• Get the days off per person and type of time off for a period with get_availability_report.
• List individual time off with filter_availability_records.
• Add time off with create_availability_record.
• Change existing time off with update_availability_record.
• Check the work days and hours a person can work in a period with get_work_schedule.
• Find the types of time off with availability_types in the list tool.

To add or change time off, the token needs write access.

An AI agent cannot delete time off. To delete time off, use ActiveCollab. See How can I delete my Availability? and How can I Delete an Availability for a different user?.

For the same data in the app, see Availability Report.