Atlassian Integration
by sooperset
mcp-atlassian connects Jira and Confluence, Cloud or Server/Data Center, to any MCP client, exposing dozens of read/write tools for issues, pages, and comments.
Add to Claude Code
claude mcp add atlassian -- uvx mcp-atlassian
About
This server gives an agent direct read and write access to Jira and Confluence: searching issues with JQL, pulling and updating ticket fields, moving issues through workflow transitions, and doing the equivalent for Confluence pages, including searching with CQL, fetching content, creating or editing pages, and posting comments. It supports both Atlassian Cloud and self-hosted Server/Data Center deployments, which matters because a lot of enterprise Jira and Confluence instances still run on-prem.
Authentication is flexible: API tokens and username/password work for Cloud and Server respectively, and OAuth 2.0 is available if you need tighter scoping. It runs as a lightweight uvx-installed process or as a Docker container behind SSE or streamable-HTTP transport, so it fits into either a local desktop MCP client or a server-side agent deployment.
Key features
- Full JQL search and CRUD on Jira issues, including workflow transitions
- CQL search plus page create/update/comment for Confluence
- Works against both Atlassian Cloud and Server/Data Center
- API token, username/password, and OAuth 2.0 authentication
- Docker image or a single uvx command, no persistent install required
- SSE and streamable-HTTP transport options for remote deployment
Use cases
- Letting an agent triage and update Jira tickets from a natural-language request
- Pulling Confluence documentation into an agent's context before it answers a support question
- Automating routine issue transitions, like closing stale tickets, as part of a larger workflow
- Bridging an on-prem Jira Server instance to a modern MCP-based agent stack
Available tools
jira_search
Runs a JQL query and returns matching issues.
jira_get_issue
Fetches the full details of a single issue by key.
jira_create_issue
Opens a new ticket with the given fields.
jira_transition_issue
Moves an issue through its workflow, e.g. To Do to In Progress to Done.
confluence_search
Runs a CQL query against Confluence and returns matching pages.
confluence_create_page
Creates a new Confluence page with the given content.
Frequently asked questions
Does this work with a self-hosted Jira/Confluence Server instance, or only Atlassian Cloud?
Both. It explicitly supports Cloud and Server/Data Center deployments, with different environment variables for each: an API token for Cloud, username/password or a personal access token for Server.
What do I need to configure before it works?
At minimum JIRA_URL and JIRA_API_TOKEN, or the Confluence equivalents, as environment variables. OAuth 2.0 is available too if you need more granular scoping.