confluence

This MCP server exposes tools to interact with Confluence from your MCP-enabled assistant. The main capabilities are: executing CQL searches to find Confluence pages, retrieving the content of a page, and updating the content of a page. You can call the tools via the MCP protocol by providing the appropriate tool name and parameters; for example, use execute_cql_search with a CQL query and an optional limit to locate relevant pages, use get_page_content with a pageId to fetch the page content, or use update_page_content with pageId, new content, and an optional new title to modify a page. The server acts as a bridge between your assistant and Confluence, returning results such as page IDs, titles, content, and any errors encountered during operations.

Prerequisites:

• Node.js and npm installed on your machine
• Basic familiarity with MCP servers and the Smithery workflow (optional but helpful)

Install dependencies and build the server:

npm install

Build the server for production or local use:

npm run build

For development with automatic rebuilds:

npm run watch

Install via Smithery (automated delivery to Claude Desktop):

npx -y @smithery/cli install @KS-GEN-AI/confluence-mcp-server --client claude

To configure Claude Desktop to use the server, add the following to the Claude Desktop config (MacOS example shown):

{
  "mcpServers": {
    "Confluence communication server": {
      "command": "node",
      "args": ["/PATH_TO_THE_PROJECT/build/index.js"],
      "env": {
        "CONFLUENCE_URL": "https://XXXXXXXX.atlassian.net/wiki",
        "CONFLUENCE_API_MAIL": "Your email",
        "CONFLUENCE_API_KEY": "KEY_FROM: https://id.atlassian.com/manage-profile/security/api-tokens"
      }
    }
  }
}

Debugging MCP servers can be challenging since they communicate over stdio. If you need debugging tools, consider using the MCP Inspector:

npm run inspector

The Inspector will provide a URL to access debugging tools in your browser.

Notes and tips:

• Ensure your Confluence credentials (URL, email, API token) are kept secure and not committed to version control.
• The build output is expected at build/index.js; adjust the path in the mcp_config if your setup differs.
• If you encounter issues with token permissions, confirm the API token has the required Confluence permissions for the actions you perform (read/write).
• When testing, start with simple calls (e.g., a basic get_page_content with a known pageId) before attempting CQL searches.
• If you modify environment variables, restart the MCP server so the changes take effect.