M
MCP Directory CURATED DIRECTORY
911 MCP servers indexed
中文
← Back to directory
M

MCP Atlassian

Community
Model Context Protocol server for Confluence and Jira
How to install ↓ Compare GitHub source repository ↗
Category
Collaboration #23 of 55
Stars
★ 6.0k Very popular
Transport
stdio (local process) · SSE · Streamable HTTP
Runtime
Python · Docker
Credentials
API key / credential required
License
MIT
Last commit
Tools
10
56FMRS · C

MCP Atlassian is a fairly comprehensive community MCP server offering roughly 98 tools for working with Jira and Confluence across Cloud and Server/Data Center deployments, with core tools covering search, CRUD on issues/pages, and status transitions — but it requires users to carefully manage the security of their own Atlassian API credentials.

Strongest · Documentation 14/20 Weakest · Reliability 7/20

Reliability
7/20
Security and permissions
9/20
Maintenance
14/20
Documentation
14/20
Setup experience
12/20
Why each score
Reliability 7/20
Only the README text is available — no manifest, source code, or tool-implementation files were provided, so the claimed '98 tools' cannot be checked against actual behavior, and the MCP init/tool-list handshake cannot be confirmed as reproducible. The README shows a 'Run Tests' badge, but no actual CI workflow file or committed test content was supplied, which falls short of the static-calibration bar requiring verifiable execution evidence. Reliability is therefore capped low, at 7, reflecting the uncertainty penalty for missing evidence.
Security and permissions 9/20
The README only broadly states 'never share API tokens' and 'keep .env files secure,' without describing scope/least-privilege controls (e.g., read-only mode, per-project/space restriction), without mentioning confirmation or safeguards for destructive operations (write/status-change tools such as jira_update_issue, jira_transition_issue, confluence_update_page), and without disclosing data-flow or external network boundaries in detail. No real tokens or secrets appear in the examples (placeholders like your_api_token are used), so this is not a red-line violation, but the unaddressed risk areas warrant a clear deduction, hence 9.
Maintenance 14/20
The repo shows PyPI version/download badges, an automated test-workflow badge, a dedicated documentation site, 5742 stars and 185 open issues, indicating meaningful adoption and attention; the MIT license is explicit and the project is not archived. However, no direct evidence of commit history, release cadence, or dependency-update frequency was provided in this review — judgment relies on indirect signals (badges/description), so it falls short of full marks, scoring 14.
Documentation 14/20
Documentation is well-layered: installation (uvx/Docker/pip/source), authentication (API token/PAT/OAuth 2.0), configuration, HTTP transport, tools reference, and troubleshooting each have dedicated linked pages, plus llms.txt/llms-full.txt for LLM consumption — showing strong documentation intent. However, the actual content lives on an external site not included in this review, so the depth of parameter-level detail and disclosed limitations could not be directly verified, only inferred from the table of contents, hence 14 rather than full marks.
Setup experience 12/20
A clear JSON config example and a one-line uvx command are provided, with a concise flow (get API token → configure IDE → use). Per the static-calibration rule, setup cannot score above 15 without verifiable execution evidence (real CI plus committed tests covering key paths); no evidence confirms the install actually succeeds, cross-platform compatibility, or diagnosable failure behavior, and the process still requires manual editing of client config files, making it a manual/fragile setup — scoring 12.

Static review · not runListed 2026-08-26

Read the FMRS scoring method →

Fit and risk

What it can accessUses the networkChanges third-party account data

Best for

  • Teams that want an AI assistant to read and write Jira issues and Confluence pages directly
  • Organizations running both Atlassian Cloud and Server/Data Center environments
  • Developers who want JQL/CQL search, issue creation, and status transitions inside their IDE or chat client

Not for

  • Users without an Atlassian Jira or Confluence account/instance
  • Scenarios requiring official Atlassian support guarantees (this is not an official Atlassian product)
  • Environments that cannot securely store or rotate API tokens/.env files

Required permissions

  • URL(s) of the Jira and/or Confluence instance
  • Atlassian account email + API token (Cloud), or a Personal Access Token (Server/Data Center)
  • Whatever read/write permissions that account actually holds in Jira/Confluence (creating/updating issues, creating/updating pages, and adding comments require write access)

Risks and side effects

  • A leaked API token or PAT can be used to read and write Jira/Confluence data as that account
  • The README explicitly warns to keep .env files secure to avoid credential exposure
  • Creating, updating, and transitioning issues, and creating/updating pages, are write operations — misconfigured automation could cause unintended data changes
  • Not officially maintained by Atlassian, so updates and security response depend on the community repository maintainers

Setup

Before you start

Runtime：Python · Docker

JIRA_URL required Your Jira instance URL, e.g. https://your-company.atlassian.net.
JIRA_API_TOKEN optionalsecret Jira API token, created at https://id.atlassian.com/manage-profile/security/api-tokens.
JIRA_PERSONAL_TOKEN optionalsecret Personal Access Token for Jira Server/Data Center deployments, used instead of username + API token.
CONFLUENCE_URL required Your Confluence instance URL, e.g. https://your-company.atlassian.net/wiki.
CONFLUENCE_API_TOKEN optionalsecret Confluence API token, created at https://id.atlassian.com/manage-profile/security/api-tokens.
Other optional settings (2)
JIRA_USERNAME optional Your Atlassian account email (for Cloud deployments).
CONFLUENCE_USERNAME optional Your Atlassian account email (for Cloud deployments).
  1. Go to https://id.atlassian.com/manage-profile/security/api-tokens and create an API token (Server/Data Center users should use a Personal Access Token instead).
  2. Add an mcp-atlassian entry to your Claude Desktop or Cursor MCP configuration, setting command to uvx, args to ["mcp-atlassian"], and env variables JIRA_URL, JIRA_USERNAME, JIRA_API_TOKEN, CONFLUENCE_URL, CONFLUENCE_USERNAME, CONFLUENCE_API_TOKEN (Server/Data Center deployments use JIRA_PERSONAL_TOKEN instead of username + API token).
  3. Alternatively, configure the same uvx-based server through Autohand Code using autohand mcp add mcp-atlassian env ....
  4. Restart your client after saving the configuration so the AI assistant can search, create, or update Jira issues and Confluence pages.
claude_desktop_config.json 
{"mcpServers": {"mcp-atlassian": {"command": "uvx", "args": ["mcp-atlassian"], "env": {"JIRA_URL": "https://your-company.atlassian.net", "JIRA_USERNAME": "[email protected]", "JIRA_API_TOKEN": "your_api_token", "CONFLUENCE_URL": "https://your-company.atlassian.net/wiki", "CONFLUENCE_USERNAME": "[email protected]", "CONFLUENCE_API_TOKEN": "your_api_token"}}}}

Shown for Claude Desktop. Other clients may use a different file or key (VS Code uses "servers") — the configurator below converts it.

.vscode/mcp.json
{
  "servers": {
    "mcp-atlassian": {
      "command": "uvx",
      "args": [
        "mcp-atlassian"
      ],
      "env": {
        "JIRA_URL": "https://your-company.atlassian.net",
        "JIRA_USERNAME": "[email protected]",
        "JIRA_API_TOKEN": "your_api_token",
        "CONFLUENCE_URL": "https://your-company.atlassian.net/wiki",
        "CONFLUENCE_USERNAME": "[email protected]",
        "CONFLUENCE_API_TOKEN": "your_api_token"
      }
    }
  }
}

Goes in your project's .vscode/mcp.json (VS Code uses a "servers" key).

Terminal
claude mcp add mcp-atlassian -e JIRA_URL=https://your-company.atlassian.net -e [email protected] -e JIRA_API_TOKEN=your_api_token -e CONFLUENCE_URL=https://your-company.atlassian.net/wiki -e [email protected] -e CONFLUENCE_API_TOKEN=your_api_token -- uvx mcp-atlassian

Run it in a terminal; replace any <…> placeholders with your own values first.

Generate and validate client config

Check that it works

After adding the config, check that tools like jira_search and confluence_get_page appear in your MCP client's tool list, then ask 'Find issues assigned to me in PROJ' to confirm the connection works.

Troubleshooting

  1. Verify JIRA_URL / CONFLUENCE_URL are correct (Cloud is typically https://your-company.atlassian.net and its /wiki path)
  2. Cloud deployments need email + API token; Server/Data Center deployments should use JIRA_PERSONAL_TOKEN instead of username + token
  3. Check whether the API token has expired or lacks sufficient permissions; regenerate it from Atlassian account security settings if needed
  4. Consult the Troubleshooting page at mcp-atlassian.soomiles.com for further debugging steps

Things to try

Once connected, you can ask your AI assistant things like:

  • Find issues assigned to me in the PROJ project
  • Search Confluence for onboarding docs
  • Create a bug ticket for the login issue
  • Update the status of PROJ-123 to Done

Tools 10

jira_search read-only
Search Jira issues using JQL (Jira Query Language)
jira_get_issue read-only
Get details for a specific Jira issue
jira_create_issue writes
Create a new Jira issue
jira_update_issue writes
Update fields on an existing Jira issue
jira_transition_issue writes
Change the workflow status of a Jira issue
confluence_search read-only
Search Confluence content using CQL (Confluence Query Language)
confluence_get_page read-only
Get the content of a specific Confluence page
confluence_create_page writes
Create a new Confluence page
Show 2 more tools
confluence_update_page writes
Update the content of an existing Confluence page
confluence_add_comment writes
Add a comment to a Confluence page

Use cases

Find issues assigned to me in the PROJ project
Search Confluence for onboarding docs
Create a bug ticket for a login issue
Update the status of PROJ-123 to Done

Supported clients

Claude Desktop
Cursor
Autohand Code

Listed from the project's documentation, not tested by this site.

Overview

MCP Atlassian is a community-maintained Model Context Protocol server that lets AI assistants interact directly with Atlassian's Jira and Confluence products. It supports both Atlassian Cloud and Server/Data Center deployments (Confluence 6.0+, Jira 8.14+), exposing 98 tools in total covering search, create, update, transition, and comment operations. The README highlights key tools such as JQL-based Jira search, issue retrieval, issue creation/update/transition, plus CQL-based Confluence search and page retrieval/creation/update/commenting. Authentication supports API tokens (Cloud), Personal Access Tokens (Server/Data Center), and OAuth 2.0 per the full documentation.

Similar servers

Compare with other MCP servers

Source revision 0a5d2427144d Data synced 2026-10-19 Read the FMRS scoring method