Castor for Confluence Data Center

Docs: https://castor.moosejudge.com/docs

A secure MCP server that bridges Atlassian Confluence Data Center (9.2 and 10.2, the two current LTS lines) with AI agents like Claude.

Castor is an independent, community-built tool. It is not affiliated with, endorsed by, or sponsored by Atlassian. Confluence® is a registered trademark of Atlassian Pty Ltd.

Overview

As Atlassian sunsets Data Center support, this server ensures organizations can still leverage AI safely within their on-premise Confluence infrastructure. It follows open standards, the MCP specification, and a "secure by default" model — see alces-mcp (the sibling Jira DC server this project mirrors) for the shared design lineage.

Features

• 7 domain tools covering spaces, pages, and CQL search, plus get_api_references for querying the generated API catalog — docs/api-catalog.md tracks the full set of cataloged Confluence DC operations and which are implemented.
• Dual-token security: CONFLUENCE_USER_TOKEN for standard operations, CONFLUENCE_ADMIN_TOKEN for elevated ones (e.g. admin_get_space_permissions) — admin tools are invisible to the agent without the admin token configured.
• Skills as MCP Resources: agent-facing operational guidance via skill://spaces-pages.

Install to use

Just want to talk to your Confluence DC instance through Claude Desktop, without cloning or building anything? Install the pre-built bundle:

1. Download the latest castor-mcp-<version>.mcpb from the GitLab Releases page.
2. In Claude Desktop: Settings → Extensions → Advanced settings → Extension Developer → "Install Extension…", then select the downloaded file.
3. Enter your Confluence URL and PAT(s) when prompted — these are stored via your OS keychain, never as plaintext.

Once connected, see docs/user/ for what you can do and how to ask for it.

If you're integrating Castor into a coding agent instead of Claude Desktop, it's also published on npm as castor-mcp — point your agent's MCP config at command: "npx", args: ["-y", "castor-mcp"] instead of cloning or building anything. See Setup to develop below only if you're contributing or need to run from an unreleased commit.

Setup to develop

1. Clone the repository.
2. Install dependencies:

   npm install

3. Copy .env.example to .env and fill in CONFLUENCE_URL and CONFLUENCE_USER_TOKEN (optionally CONFLUENCE_ADMIN_TOKEN).
4. Build the project:

   npm run build

5. Run tests:

   npm test

There is no running local Confluence instance yet — docker-compose.yml defines one (see docs/developer/testing-environment.md), but it requires a Confluence developer license this project does not currently have, and has never been started.

Documentation

• docs/user/ — task-oriented guide to what the agent can do.
• docs/developer/ — codebase notes for contributors.
• docs/api-catalog.md — generated reference of every cataloged Confluence DC operation and its implementation status.
• docs/specs/, docs/plans/ — design specs and implementation plans.