atlassian-agent¶

Comprehensive MCP server + A2A agent for Jira and Confluence management across Atlassian Cloud and Server.

Official documentation

This site is the canonical reference for atlassian-agent, maintained alongside every release.

Overview¶

atlassian-agent wraps the Atlassian REST surface — Jira and Confluence, for both Cloud and Server / Data Center — with typed, deterministic MCP tools, and ships a Pydantic-AI A2A agent that drives those tools conversationally. It provides:

A broad Atlassian tool surface — Jira projects, issues, comments, fields, screens, workflows, and users, alongside Confluence pages, spaces, and users, plus Cloud organization administration (org, admin, API access, DLP, user management, user provisioning, control).
Two console scripts — atlassian-mcp (the MCP server) and atlassian-agent (the A2A agent server with an optional web UI).
Tolerant REST clients — every call returns a structured Response rather than raising, so the surface remains usable without a reachable instance and remains inactive when credentials are absent.

Explore the documentation¶

Installation — pip, source, extras, and the prebuilt Docker image.
Deployment — run the MCP server, the agent server, Docker Compose, Caddy + Technitium.
Usage — the MCP tools, the Atlassian Python clients, and the CLI.
Architecture — the layered tool / API / agent design.
Concepts — the CONCEPT:ATL-* registry.

Quick start¶

pip install "atlassian-agent[mcp]"
atlassian-mcp                       # stdio MCP server (default transport)

Connect it to an Atlassian instance:

export ATLASSIAN_AGENT_URL=https://your-company.atlassian.net
export ATLASSIAN_AGENT_USER=your-email@example.com
export ATLASSIAN_AGENT_TOKEN=your_api_token
atlassian-mcp --transport streamable-http --host 0.0.0.0 --port 8000

See Installation and Deployment for the full matrix (PyPI extras, Docker image, all transports, the agent server, reverse proxy, DNS).