Asana MCP Server

What is Asana MCP Server?

The Asana MCP Server acts as a bridge between Large Language Model (LLM) clients and the Asana API, providing a uniform and secure interface for accessing, searching, creating, updating, and analyzing Asana workspaces, projects, tasks, comments, and statuses. With read-only and full-access modes, it adapts to a range of security and workflow requirements.

How to Configure Asana MCP Server

1. Create an Asana account:

   • Sign up at Asana.

2. Generate an Asana Access Token:

   • Visit Asana Developer Console and create a Personal Access Token.
   • More details: Asana Token Guide

3. Install the Asana MCP Server:

   • For integration with Claude Desktop, add the following configuration to your claude_desktop_config.json:

     {
       "mcpServers": {
         "asana": {
           "command": "npx",
           "args": ["-y", "@roychri/mcp-server-asana"],
           "env": {
             "ASANA_ACCESS_TOKEN": "your-asana-access-token"
           }
         }
       }
     }
     

   • For Claude Code, run:

     claude mcp add asana -e ASANA_ACCESS_TOKEN=<TOKEN> -- npx -y @roychri/mcp-server-asana
     

     Replace <TOKEN> with your Asana Access Token.

4. Environment Variables:

   • ASANA_ACCESS_TOKEN (required): Your Asana API token.
   • READ_ONLY_MODE (optional): Set to true to disable write operations, ideal for testing or read-only workflows.

5. Advanced:

   • To test changes locally, use the included MCP Inspector.
   • For beta versions, use @roychri/mcp-server-asana@beta.

How to Use Asana MCP Server

Once configured and running, the Asana MCP Server automatically makes Asana tools and resources available to your MCP-enabled AI application (e.g., Claude Desktop). To use it:

1. Simply ask natural language queries about Asana workspaces, projects, or tasks.
2. For more targeted interactions, mention "asana" in your prompt to help the LLM select the correct tool.
3. The server exposes dozens of callable tools such as searching for tasks, creating projects, listing comments, and analyzing task completeness.
4. You can monitor and test tool invocations via your AI application's UI or the MCP Inspector.
5. In read-only mode, only safe operations are enabled; attempts to create or modify data are blocked.

Key Features

• Comprehensive Integration: Direct and secure linking of AI applications to the Asana API, covering most core functionalities.
• Rich Toolset: Over 20 callable tools for searching, creating, updating, and analyzing tasks, projects, statuses, and comments.
• Customizable Access: Supports both read-only and read-write operation modes for enhanced security and control.
• Prompt Templates: Built-in prompt generators for common scenarios like status summaries and task completeness analysis.
• Resource Representation: Full exposure of Asana workspaces and projects as structured MCP resources for rich context building.
• Scalable Architecture: Designed for both local and cloud/server deployment.
• Open Source & MIT License: Free to use, modify, and extend in any context.

Use Cases

• Automated Project Reporting: Have your AI summarize project or task status, uncover bottlenecks, or surface overdue work items.
• Smart Task Management: Interact with Asana via natural language to create, update, or search for tasks hands-free.
• Workflow Automation: Power agents that trigger Asana updates based on conversations or other AI-driven workflows.
• Project Audit & Analysis: Use built-in tools to check for task completeness, extract project stats, or analyze work distribution.
• Testing and Read-only Monitoring: Safely explore or monitor your Asana environment in read-only mode for compliance or demonstrations.

FAQ

Q1: What permissions does my Asana account need to use the MCP Server?
A1: You must have an Asana plan (free or paid) that allows API access, and your generated Access Token must permit the required read or write operations for your use case.

Q2: How do I run the Asana MCP Server in read-only mode to prevent accidental changes?
A2: Set the environment variable READ_ONLY_MODE=true during server startup. Creation, update, and delete tools will be disabled.

Q3: The server returns permission errors. How do I resolve this?
A3: Double-check that your Asana Access Token is correct and active. Ensure your Asana account has API access, and the server configuration matches your credentials in claude_desktop_config.json.

Q4: How do I test the server locally without a full deployment?
A4: Use the MCP Inspector by running npm run inspector in the server directory. You can specify custom ports with CLIENT_PORT and SERVER_PORT if needed.

Q5: Can I use this server with LLMs other than Claude?
A5: Yes, as long as your AI tool supports the Model Context Protocol, it can connect to the Asana MCP Server.