Skip to navigation
Getting started

MCP Setup

The ProjectManager MCP Server lets AI assistants such as Claude work directly with your ProjectManager.com workspace. Once connected, you can ask your assistant questions about your projects and tasks in plain language, and have it make updates for you, without writing any code.

What the MCP Server does

The Model Context Protocol (MCP) is an open standard that lets AI assistants connect to external tools and data sources. The ProjectManager MCP Server exposes your workspace to any MCP-compatible client as a set of tools the assistant can call on your behalf.

With the MCP Server connected, your assistant can:

  • Look up projects and tasks – find projects, review task lists, and check what is assigned to you or what is overdue.
  • Summarize project health – pull project summaries, metrics, and planned progress to answer questions like “Which of my projects are behind schedule?”
  • Review team workload – see resources, project members, and their workload to help balance assignments.
  • Create and update work – create projects and tasks, change task status and priority, add or remove assignees and tags, and organize tasks under parent tasks.
  • Read discussions – retrieve the comments on a task to catch up on the conversation.

The MCP Server uses the same permissions as the API key it connects with. The assistant can only see and change what that key has access to, and any changes it makes appear in your workspace’s audit logs just like changes made through the API.

General setup

Any MCP client that supports remote (Streamable HTTP) servers can connect to ProjectManager. You need three things: the right role permission, an API key, and the MCP Server URL.

1. Check your role permission

To use the MCP Server, your role must include the Access work from ChatGPT and Claude permission. This permission is turned on by default for the Admin and Global Admin roles.

To give this permission to another role, a global admin can open the roles screen in ProjectManager, select the role, and enable Access work from ChatGPT and Claude.

2. Generate an API key

  1. API keys are tied to workspaces, and are available to users with admin or global admin roles only.
  2. Click on your name in the bottom left hand corner, then select “Integrations”.
  3. On the “Integrations” screen, select the “API Keys” tab on the left hand side.
  4. Click “Create API Key”.
  5. Give your API key a name that you can use to remember it, for example “Claude MCP”.
  6. Click “Generate Tokens” then copy and save your API key.
  7. We recommend rotating API keys - follow your organization’s key rotation policy.

See Authentication for more details on API keys.

3. Configure your MCP client

Add a new remote MCP server to your client with the following settings:

  • MCP Server URL: https://api.projectmanager.com/api/mcp
  • Transport: Streamable HTTP
  • Authentication: Bearer token, using the API key you generated

Clients that let you set request headers directly should send the API key using the Authorization header:

Authorization: Bearer <your-api-key>

Your API key grants access to your workspace. Treat it like a password: don’t share it, don’t commit it to source control, and revoke it from the “API Keys” tab if you think it has been exposed.