mcp-atlassian

MCP Atlassian Server connects AI agents like Cline, Claude Desktop, or Cursor to Atlassian Jira and Confluence, exposing a standardized set of resources and tools that let AI agents read data (resources) and perform actions (tools) against Jira and Confluence. The server is designed to be local-first and developer-friendly, with easy integration for Cline via MCP Marketplace while remaining compatible with other MCP clients. Once running, you can query Jira resources such as issues, projects, and users, browse Confluence spaces and pages, and invoke tools to create or update issues, manage pages, or handle comments. The system abstracts Atlassian API interactions behind MCP resources and tools, enabling seamless AI-assisted workflows and automation across Atlassian products.

To use it, install the MCP Atlassian Server in your development or deployment environment (via Smithery or your preferred MCP client setup). After starting the server, configure your MCP client to route appropriate queries to jira:// or confluence:// endpoints exposed by the server. You can leverage the available Jira and Confluence tools to perform mutations like creating issues or pages, updating statuses, or adding comments, as well as read-only resources for discovery and reporting.

Prerequisites:

• Node.js (recommended LTS) and npm installed on your machine
• Basic familiarity with MCP concepts and your Atlassian API tokens

Installation steps:

1. Install Node.js if not already installed. Visit https://nodejs.org/ and follow the installer for your OS.

2. Install the MCP Atlassian Server via Smithery (recommended) or directly use npx to install and run:

Using Smithery:

npx -y @smithery/cli install @phuc-nt/mcp-atlassian-server --client claude

This will install and configure the server for the Claude Desktop client. You can choose another client or run without Smithery if desired.

3. Alternatively, run directly with npm/npx (no Smithery):

npx -y @phuc-nt/mcp-atlassian-server

4) Configure Atlassian API credentials:
- Generate and securely store your Jira and Confluence API tokens.
- Provide necessary environment variables or configuration files as documented by the server (e.g., ATLASSIAN_API_TOKEN, ATLASSIAN_BASE_URL, etc.).

5) Start the server and connect your MCP client:
- Ensure your MCP client is configured to communicate with the MCP server endpoints exposed by this package.
- Follow the client-specific setup instructions to register the resources and tools provided by the MCP Atlassian Server.

Notes:
- The installation guide llms-install.md is optimized for AI assistants (like Cline) to automatically set up the server. If you’re not using Cline, refer to the docs for standard MCP client integration.
- For security, store API tokens securely and avoid embedding secrets in code or configuration files that may be shared.

Tips and troubleshooting:

• Ensure your Atlassian tokens have the minimum required scopes for the resources you intend to access (issues, boards, pages, attachments, comments, etc.).
• If you see connectivity errors, verify network access from the MCP server to Jira/Confluence endpoints and check any firewall rules.
• When upgrading, review CHANGELOG.md for breaking changes related to resource and tool availability or API compatibility.
• Enable verbose logs during initial setup to diagnose authentication or permission issues more effectively.
• If you plan to deploy in a production environment, consider containerization (Docker) and secure secret management for API tokens.