# atlassian-confluence-mcp-server [Health: Active]

**Category:** 🧠 Knowledge & Memory  
**Repository:** https://github.com/anuragagrawal0430/atlassian-confluence-mcp-server  
**GitHub Stars:** 5  
**npm Downloads (last month):** 1267  
**Views:** 0  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/atlassian-confluence-mcp-server

## Description
Model Context Protocol (MCP) server for Atlassian Confluence. Enables AI assistants like Claude and Windsurf to read, create, search, and manage wiki pages on Cloud, on-premise Server, and Data Center.

## Tools
Capabilities this server exposes over MCP:

- **confluence_test_connection** — Test connection to Confluence API
- **confluence_get_spaces** — List all spaces
- **confluence_get_space** — Get space details by key
- **confluence_get_space_by_key** — Alias for `get_space
- **confluence_create_space** — Create a new space
- **confluence_delete_space** — Delete a space and all its content
- **confluence_get_space_homepage** — Get the homepage of a space
- **confluence_get_pages** — List pages, optionally filtered by space
- **confluence_get_page** — Get page by ID
- **confluence_get_page_by_title** — Get page by title within a space
- **confluence_create_page** — Create a new page
- **confluence_update_page** — Update an existing page
- **confluence_patch_page** — Server-side exact find/replace patch for large pages
- **confluence_replace_page_range** — Replace middle range using offsets or line numbers
- **confluence_append_to_page** — Append or prepend content server-side
- **confluence_get_page_body_chunk** — Read large page body in offset/length chunks
- **confluence_delete_page** — Delete a page
- **confluence_copy_page** — Copy a page, optionally to a different space
- **confluence_move_page** — Move a page to a different parent or space
- **confluence_search** — Search using CQL
- **confluence_search_pages** — Search pages by text
- **confluence_get_recently_modified** — Get recently modified pages
- **confluence_get_page_labels** — Get labels on a page
- **confluence_add_page_label** — Add a label to a page
- **confluence_delete_page_label** — Remove a label from a page
- **confluence_get_page_comments** — Get comments on a page
- **confluence_add_page_comment** — Add a comment to a page
- **confluence_get_page_attachments** — List attachments on a page
- **confluence_get_comment_attachments** — List attachments on a comment
- **confluence_upload_attachment** — Upload/upsert a local file as a page or comment attachment
- **confluence_get_child_pages** — Get child pages of a parent
- **confluence_get_page_ancestors** — Get ancestor pages
- **confluence_get_page_versions** — Get page version history
- **confluence_get_page_version** — Get a specific page version
- **confluence_get_page_properties** — Get content properties of a page
- **confluence_get_current_user** — Get current authenticated user
- **confluence_get_page_watchers** — Get users watching a page
- **confluence_watch_page** — Add current user as a watcher
- **confluence_unwatch_page** — Remove current user as a watcher
- **confluence_get_page_permissions** — Get page restrictions/permissions
- **confluence_set_page_permissions** — Set page restrictions/permissions
- **confluence_get_page_tasks** — Extract inline tasks from a page
- **confluence_export_page** — Get export URL for a page (PDF or Word)
- **confluence_get_page_by_url** — Get page from a Confluence URL
- **confluence_get_personal_space_key** — Get current user's personal space key
- **confluence_create_page_in_personal_space** — Create page in personal space
- **confluence_create_private_space** — Create a private space

## Claude Desktop Quick Installation
Install path detected from listing signals. Uses `npx` (confidence: high):

```json
"mcpServers": {
  "atlassian-confluence-mcp-server": {
    "command": "npx",
    "args": ["-y","atlassian-confluence-mcp-server"],
    "env": {
      "CONFLUENCE_BASE_URL": "",
      "PAT": "",
      "CONFLUENCE_USERNAME": "",
      "CONFLUENCE_PASSWORD": ""
    }
  }
}
```

**Requires environment variables:** `CONFLUENCE_BASE_URL`, `PAT`, `CONFLUENCE_USERNAME`, `CONFLUENCE_PASSWORD` — the values above are empty placeholders; fill in real credentials before running (see the repository for what each one is for).

## Documentation

## What atlassian-confluence-mcp-server does

atlassian-confluence-mcp-server exposes Confluence operations through the Model Context Protocol. It supports Confluence Cloud as well as on-premise Server and Data Center deployments running version 7.9 or later. The package includes 47 tools covering content retrieval, page and space administration, search, collaboration features, and selected export operations.

Its coverage includes page creation, reading, updating, deletion, copying, and movement. Agents can also work with spaces, labels, comments, attachments, page hierarchies, version history, watchers, page restrictions, inline tasks, personal-space pages, and Confluence URLs. Search supports both text queries and CQL, along with recently modified pages. Export tools generate PDF or Word export URLs.

## How it works

The server runs as a standard command-line MCP process. It can be started directly with `npx atlassian-confluence-mcp-server`, installed globally, or built from the repository and launched with Node.js. MCP client configuration supplies the command and environment variables, allowing an assistant to invoke Confluence tools through the server.

Authentication depends on the Confluence deployment. Cloud uses an Atlassian account email together with an API token. Server and Data Center use a Personal Access Token with Bearer authentication; older Server installations can use username and password authentication instead.

Read-only operation is the default. Mutating tools require `CONFLUENCE_READ_ONLY=false`, while deletion and page-permission mutation require the separate `CONFLUENCE_ENABLE_DESTRUCTIVE_TOOLS=true` setting. Tool and space allowlists provide additional ways to narrow what the process can expose or access.

## Setup and configuration

Node.js 22.12 or newer is required. Set `CONFLUENCE_BASE_URL` to the root URL of the target Confluence instance. Provide `PAT` for a Cloud API token or a Server/Data Center personal access token. Cloud authentication also requires `CONFLUENCE_USERNAME` with the Atlassian account email. Basic authentication uses `CONFLUENCE_USERNAME` and `CONFLUENCE_PASSWORD`.

Optional controls include `CONFLUENCE_ENABLED_TOOLS`, a comma-separated tool allowlist, and `CONFLUENCE_ALLOWED_SPACES`, which limits requests using space keys. HTTPS is required by default; insecure HTTP can only be enabled for localhost or loopback addresses. Attachment uploads are disabled unless `CONFLUENCE_UPLOAD_ALLOWED_DIRS` names permitted absolute directories. `CONFLUENCE_MAX_ATTACHMENT_BYTES` limits the size of files read for upload and defaults to 50 MB.

A typical MCP entry uses `npx -y atlassian-confluence-mcp-server` as the command and places the Confluence settings in its `env` object. The same pattern works with the globally installed command or a built `dist/index.js` entry point.

## Tools and capabilities

The available tool set covers:

- Confluence connection testing
- Space listing, inspection, creation, and deletion
- Page CRUD, copying, moving, hierarchy navigation, and URL parsing
- CQL and text search, including recently changed pages
- Labels, comments, attachments, watchers, and version history
- Page restrictions and permission management
- Inline task extraction and PDF or Word export URLs

The effective tool set depends on the security settings. Read-only mode hides or blocks mutation operations, and an explicit enabled-tools list can expose only selected functions.

## Limitations and notes

The server requires access to a reachable Confluence instance and credentials with suitable permissions. Cloud and Server/Data Center authentication use different credential patterns. Confluence may impose an attachment limit below the client-side upload cap. File uploads read from the host filesystem only when an allowed directory has been configured. Destructive operations remain disabled unless they are explicitly enabled.

_Full upstream README: https://allmcps.com/mcp/atlassian-confluence-mcp-server/readme_