mcp-atlassian

What it is and what it does

mcp-atlassian is a Model Context Protocol server that bridges Atlassian products (Jira and Confluence) with AI language models. It allows LLMs like Claude to query, create, and update Jira issues and Confluence pages through a standardized MCP interface, supporting both Atlassian Cloud and on-premises Server/Data Center deployments. The server exposes tools for searching with JQL (Jira Query Language) and CQL (Confluence Query Language), retrieving and modifying issues and pages, transitioning workflows, and adding comments.

The package is designed to integrate with AI development environments (Claude Desktop, Cursor) and Autohand Code, enabling natural-language workflows like "find issues assigned to me" or "search Confluence for onboarding docs." It handles authentication via API tokens or Personal Access Tokens, manages HTTP transport over SSE or streamable-http, and includes markdown-to-Confluence conversion for seamless documentation updates. With 27 runtime dependencies including uvicorn, httpx, and the atlassian-python-api client, it provides a production-ready bridge for AI-driven Atlassian automation.

Use it for:

• Ask an AI assistant to search and retrieve Jira issues by JQL query, then summarize or analyze them without manual dashboard navigation.
• Automate Confluence documentation updates by having an LLM generate markdown content and push it directly to pages via the MCP server.
• Create and transition Jira tickets from natural-language descriptions—e.g., 'create a bug for the login issue and assign it to the backend team.'
• Build AI-powered onboarding or knowledge-base workflows that search Confluence for relevant docs and surface them in conversation context.
• Integrate Atlassian operations into multi-step LLM workflows, such as fetching issue details, generating summaries, and posting results as Confluence comments.

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

An MCP server that exposes Jira and Confluence APIs to AI language models, enabling LLM-driven queries, issue management, and documentation workflows across Atlassian Cloud and Server/Data Center deployments.

Yes, if you use Atlassian products (Jira or Confluence) and want to integrate them with AI assistants or LLM workflows. The package is actively maintained, has low install friction, and offers broad tool coverage. However, verify the actual license terms before use in proprietary contexts, since the metadata does not record the license formally. Not relevant if you do not use Atlassian or do not need LLM-driven automation.

Install

mcp-atlassian on PyPI

pip install mcp-atlassian

uv add mcp-atlassian

poetry add mcp-atlassian

Installing mcp-atlassian

Before you install

Low install friction with a pure-Python wheel distribution. Active maintenance (latest release 2026-07-18, 27 days ago). Requires Python 3.10 or later and pulls in 27 runtime dependencies including web frameworks (uvicorn, starlette), HTTP clients (httpx, requests), and Atlassian API bindings.

License in practice

License treatment is unclear—no SPDX identifier or raw license text is recorded in the package metadata, despite the description mentioning MIT licensing. Verify the actual license terms before use in proprietary or restricted contexts.

Quickstart

# Install
pip install mcp-atlassian

# Configure environment
export JIRA_URL="https://your-company.atlassian.net"
export JIRA_USERNAME="your.email@company.com"
export JIRA_API_TOKEN="your_api_token"
export CONFLUENCE_URL="https://your-company.atlassian.net/wiki"
export CONFLUENCE_USERNAME="your.email@company.com"
export CONFLUENCE_API_TOKEN="your_api_token"

# Run as MCP server (typically via uvx or IDE integration)
uvx mcp-atlassian

Requires Python 3.10 or later. Atlassian API credentials (Cloud API token or Server/Data Center Personal Access Token) must be set as environment variables before the server starts. Designed to run as a long-lived MCP server process, not as a library for direct Python calls.

Verify before relying

• Whether the package actually implements all 93 tools mentioned in the description or if that count includes planned/partial implementations.
• Support scope for Server/Data Center versions (v6.0+ for Confluence, v8.14+ for Jira) and whether older versions are tested.
• Whether OAuth 2.0 authentication is fully implemented or documented as in-progress.
• Performance characteristics and rate-limit handling when querying large Jira/Confluence instances.

Package facts

Evidence: mcp_atlassian-0.23.0-py3-none-any.whl

Tags

More WWW/HTTP packages

Further reading