Confluence

Confluence MCP Server provides a standardized interface to interact with Confluence content for AI assistants. It handles authentication, retrieval, creation and updates of pages, comments, and attachments, and it formats data for AI consumption while offering basic rate limiting to protect the API.

How to use

You run the Confluence MCP Server locally or in your environment and connect to it from your MCP client. The server exposes tools to retrieve pages, search content, list spaces, create and update pages, manage comments and attachments, and transform Confluence content for AI consumption. Use these capabilities to power AI-assisted queries, content updates, or automated workflows against your Confluence instance.

How to install

Prerequisites you need before installing the MCP server are Bun (v1.0.0 or higher) and a Confluence account with API access.

# Clone the repository
git clone https://github.com/cosmix/confluence-mcp.git
cd confluence-mcp

# Install dependencies
bun install

# Build the project
bun run build

Configuration and setup

To run the MCP server you must provide your Confluence credentials and base URL through environment variables.

CONFLUENCE_API_TOKEN=your_api_token
CONFLUENCE_BASE_URL=your_confluence_instance_url/wiki
CONFLUENCE_USER_EMAIL=your_email

Run the server from your MCP client

The server is started using a local runtime. You specify the path to the built index file and any required environment variables.

{
  "type": "stdio",
  "name": "confluence",
  "command": "bun",
  "args": ["/absolute/path/to/confluence-mcp/dist/index.js"],
  "env": {
    "CONFLUENCE_API_TOKEN": "your_api_token",
    "CONFLUENCE_BASE_URL": "your_confluence_instance_url/wiki",
    "CONFLUENCE_USER_EMAIL": "your_email"
  }
}

Accessing tools and common actions

The Confluence MCP Server exposes a suite of tools to work with Confluence content. You can retrieve pages by ID, search pages with CQL, list spaces, create and update pages, as well as manage comments and attachments. Each action is designed to be used by your MCP client to perform specific tasks against your Confluence instance.

Notes on security and usage

Keep your API token secure. Do not share your environment variables or include them in logs. If you rotate tokens, update the running MCP server configuration accordingly. The server includes basic rate limiting to prevent API abuse.

Available tools

get_page

Retrieve a Confluence page by ID in a chosen format (text, markdown) with optional original XHTML markup for updates.

search_pages

Search Confluence pages using Confluence Query Language (CQL) and return content in a chosen format.

get_spaces

List all available Confluence spaces with optional limits.

create_page

Create a new Confluence page in a specified space, optionally under a parent page.

update_page

Update an existing Confluence page with a new title and content and version control.

get_comments

Retrieve comments for a specific Confluence page in a chosen format.

add_comment

Add a new comment to a Confluence page, with optional threaded reply via parent ID.

get_attachments

Retrieve attachments for a Confluence page.

add_attachment

Add an attachment to a Confluence page, including optional changelog/comment.