Get the FREE Ultimate OpenClaw Setup Guide →

atlassian

Remote MCP Server that securely connects Jira and Confluence with your LLM, IDE, or agent platform of choice.

Installation
Run this command in your terminal to add the MCP server to Claude Code.
Run in terminal:
Command
claude mcp add --transport stdio atlassian-atlassian-mcp-server node server.js \
  --env PORT="Port to run the MCP server on" \
  --env SCOPES="OAuth scopes required for Jira, Compass, Confluence access" \
  --env CLOUD_ID="Your Atlassian Cloud site ID" \
  --env ATLAS_API_TOKEN="Optional API token for headless authentication" \
  --env ATLAS_CLIENT_ID="OAuth2.1 client ID for Atlassian" \
  --env ATLAS_CLIENT_SECRET="OAuth2.1 client secret for Atlassian"

How to use

The Atlassian Rovo MCP Server acts as a cloud-based bridge between your Atlassian Cloud instance (Jira, Compass, and Confluence) and external tools and AI assistants. It enables tools to search, summarize, create, and update Atlassian data in real time while honoring existing permissions through OAuth 2.1 or API tokens. Once configured, supported clients such as OpenAI ChatGPT, Claude, Copilot CLI, Gemini CLI, Amazon Quick Suite, and VS Code can connect to the MCP server to perform tasks like querying Jira issues, summarizing Confluence pages, or creating new content based on natural language prompts. The server also supports any local MCP-compatible client via the mcp-remote proxy, enabling custom integrations that follow the MCP specification.

To use it, deploy the MCP server and connect a client from the supported list. The client will authenticate with OAuth 2.1 or an API token, establish a streaming session, and then issue context-aware requests such as searching Jira issues, summarizing a Confluence page, or creating a new Jira ticket from meeting notes. Each action respects Atlassian permissions and the configured scopes, ensuring secure access control across your Atlassian data.

How to install

Prerequisites

  • Node.js v18+ (for running the MCP server and the mcp-remote proxy)
  • npm or yarn
  • An Atlassian Cloud site with Jira, Compass, and/or Confluence
  • Optional: OAuth 2.1 client credentials or an API token for headless authentication

Install steps

  1. Clone or download the MCP server repository
  2. Install dependencies
npm install
  1. Configure environment variables Create a .env file (or set environment variables in your hosting environment) with entries such as:
PORT=3000
ATLAS_CLIENT_ID=your-oauth-client-id
ATLAS_CLIENT_SECRET=your-oauth-client-secret
ATLAS_API_TOKEN=your-api-token (optional if using OAuth)
CLOUD_ID=your-cloud-id
SCOPES=required OAuth scopes for Jira, Confluence, Compass
  1. Run the server
node server.js
  1. If you plan to expose a client via mcp-remote, install and run the client proxy per its documentation and point it to the MCP server URL (e.g., https://your-domain.com/v1/mcp).

Additional notes

Tips and common issues:

  • Ensure OAuth 2.1 scopes align with the data you want to access (Jira, Confluence, Compass). Misconfigured scopes will block access.
  • If using API tokens for headless clients, admin enablement is required on the Atlassian side.
  • Use HTTPS in production; TLS 1.2 or newer is recommended as noted in the security guidance.
  • When updating custom clients, point them to the /mcp endpoint rather than legacy /sse endpoints as recommended by Atlassian.
  • If you encounter permission errors, verify the user or token has access to the specific Jira projects, Confluence spaces, or Compass components involved.
  • For local testing with mcp-remote, ensure Node.js v18+ is installed and the local URL is reachable from your client.

Related MCP Servers

Sponsor this space

Reach thousands of developers