Transform how you access and interact with your team's knowledge by connecting Claude, Cursor AI, and other AI assistants directly to your Confluence spaces, pages, and documentation. Get instant answers from your knowledge base, search across all your spaces, and streamline your documentation workflow.
Node.js/TypeScript MCP server for Atlassian Confluence. Provides tools enabling AI systems (LLMs) to list/get spaces & pages (content formatted as Markdown) and search via CQL. Connects AI seamlessly to Confluence knowledge bases using the standard MCP in
2026-11-17
星數
—
工具
—
可用率
—
What You Can Do
Ask AI about your documentation: "What's our API authentication process?"
Search across all spaces: "Find all pages about security best practices"
Get instant answers: "Show me the latest release notes from the Product space"
Access team knowledge: "What are our HR policies for remote work?"
Review page comments: "Show me the discussion on the architecture document"
Create and update content: "Create a new page in the DEV space"
Perfect For
Developers who need quick access to technical documentation and API guides
Product Managers searching for requirements, specs, and project updates
HR Teams accessing policy documents and employee resources quickly
Support Teams finding troubleshooting guides and knowledge base articles
Anyone who wants to interact with Confluence using natural language
Copy the generated token immediately (you won't see it again!)
2. Try It Instantly
# Set your credentials
export ATLASSIAN_SITE_NAME="your-company" # for your-company.atlassian.net
export ATLASSIAN_USER_EMAIL="[email protected]"
export ATLASSIAN_API_TOKEN="your_api_token"
# List your Confluence spaces (TOON format by default)
npx -y @aashari/mcp-server-atlassian-confluence get --path "/wiki/api/v2/spaces"
# Get details about a specific space with field filtering
npx -y @aashari/mcp-server-atlassian-confluence get \
--path "/wiki/api/v2/spaces/123456" \
--jq "{id: id, key: key, name: name, type: type}"
# Get a page with JMESPath filtering
npx -y @aashari/mcp-server-atlassian-confluence get \
--path "/wiki/api/v2/pages/789" \
--jq "{id: id, title: title, status: status}"
# Search for pages (using CQL)
npx -y @aashari/mcp-server-atlassian-confluence get \
--path "/wiki/rest/api/search" \
--query-params '{"cql": "type=page AND space=DEV"}'
Connect to AI Assistants
For Claude Desktop Users
Add this to your Claude configuration file (~/.claude/claude_desktop_config.json):
Then configure your AI assistant to use the MCP server with STDIO transport. The binary is available as mcp-atlassian-confluence after global installation.
Alternative: Configuration File
Create ~/.mcp/configs.json for system-wide configuration:
Alternative config keys: The system also accepts "atlassian-confluence", "@aashari/mcp-server-atlassian-confluence", or "mcp-server-atlassian-confluence" instead of "confluence".
Using Environment Variables
You can also configure credentials using environment variables or a .env file:
# Create a .env file in your project directory
cat > .env << EOF
ATLASSIAN_SITE_NAME=your-company
[email protected]
ATLASSIAN_API_TOKEN=your_api_token
DEBUG=false
EOF
The server will automatically load these values from:
Environment variables
.env file in the current directory
~/.mcp/configs.json (as shown above)
Available Tools
This MCP server provides 5 generic tools that can access any Confluence API endpoint:
Tool
Description
conf_get
GET any Confluence API endpoint (read data)
conf_post
POST to any endpoint (create resources)
conf_put
PUT to any endpoint (replace resources)
conf_patch
PATCH to any endpoint (partial updates)
conf_delete
DELETE from any endpoint (remove resources)
Tool Parameters
All tools share these common parameters:
path (required): The API endpoint path (e.g., /wiki/api/v2/spaces)
What is TOON? TOON (Token-Oriented Object Notation) is a format optimized for LLM token efficiency, reducing token costs by 30-60% compared to JSON. It's the default output format for all tools.
Benefits:
Tabular arrays use fewer tokens than JSON arrays
Minimal syntax overhead (no quotes, brackets, commas where unnecessary)
Still human-readable and parseable
When to use JSON instead:
When you need standard JSON for other tools
When debugging or manual inspection is needed
Example comparison:
// JSON format (verbose)
{"results": [{"id": "123", "title": "My Page"}, {"id": "456", "title": "Other Page"}]}
// TOON format (efficient)
results:
- id: 123
title: My Page
- id: 456
title: Other Page
To use JSON instead of TOON, set outputFormat: "json" in your request.
JMESPath Filtering
All tools support optional JMESPath (jq) filtering to extract specific data and reduce token costs:
# Get just space names and keys
npx -y @aashari/mcp-server-atlassian-confluence get \
--path "/wiki/api/v2/spaces" \
--jq "results[].{id: id, key: key, name: name}"
# Get page title and status
npx -y @aashari/mcp-server-atlassian-confluence get \
--path "/wiki/api/v2/pages/123456" \
--jq "{id: id, title: title, status: status}"
IMPORTANT: Always use the jq parameter to filter responses to only the fields you need. Unfiltered responses can be very large and expensive in token costs.
-b, --body <json> (required) - Request body as JSON
Examples
# GET request
npx -y @aashari/mcp-server-atlassian-confluence get --path "/wiki/api/v2/spaces"
# GET with query parameters and JMESPath filter
npx -y @aashari/mcp-server-atlassian-confluence get \
--path "/wiki/api/v2/pages" \
--query-params '{"space-id": "123456", "limit": "10"}' \
--jq "results[*].{id: id, title: title}"
# GET with JSON output format
npx -y @aashari/mcp-server-atlassian-confluence get \
--path "/wiki/api/v2/spaces" \
--output-format json
# POST request (create a page)
npx -y @aashari/mcp-server-atlassian-confluence post \
--path "/wiki/api/v2/pages" \
--body '{"spaceId": "123456", "status": "current", "title": "New Page", "body": {"representation": "storage", "value": "<p>Content here</p>"}}'
# POST request (add a comment)
npx -y @aashari/mcp-server-atlassian-confluence post \
--path "/wiki/api/v2/pages/789/footer-comments" \
--body '{"body": {"representation": "storage", "value": "<p>My comment</p>"}}'
# PUT request (update page - requires version increment)
npx -y @aashari/mcp-server-atlassian-confluence put \
--path "/wiki/api/v2/pages/789" \
--body '{"id": "789", "status": "current", "title": "Updated Title", "spaceId": "123456", "body": {"representation": "storage", "value": "<p>Updated content</p>"}, "version": {"number": 2}}'
# PATCH request (partial update)
npx -y @aashari/mcp-server-atlassian-confluence patch \
--path "/wiki/api/v2/spaces/123456" \
--body '{"name": "New Space Name"}'
# DELETE request
npx -y @aashari/mcp-server-atlassian-confluence delete \
--path "/wiki/api/v2/pages/789"
Response Handling
Large Response Truncation
When API responses exceed approximately 40,000 characters (~10,000 tokens), the server automatically truncates the response to stay within token limits. When this happens:
You'll see a truncation notice at the end of the response showing:
How much of the original response is shown
The original response size
Guidance on accessing the full data
The full raw response is saved to a temporary file in /tmp/mcp/ (path provided in the truncation notice)
Best practices to avoid truncation:
Always use the jq parameter to filter responses to only needed fields
Use limit query parameter to restrict result counts (e.g., {"limit": "5"})
Request specific resources by ID rather than listing all
Use targeted CQL queries for searches
Example of efficient filtering:
# Instead of getting all space data (can be huge):
npx -y @aashari/mcp-server-atlassian-confluence get \
--path "/wiki/api/v2/spaces"
# Get only the fields you need:
npx -y @aashari/mcp-server-atlassian-confluence get \
--path "/wiki/api/v2/spaces" \
--query-params '{"limit": "10"}' \
--jq "results[*].{id: id, key: key, name: name}"
Debug Logging
Enable debug logging to see detailed request/response information:
# Set DEBUG environment variable
export DEBUG=true
# For MCP mode
DEBUG=true npx -y @aashari/mcp-server-atlassian-confluence
# For CLI mode
DEBUG=true npx -y @aashari/mcp-server-atlassian-confluence get --path "/wiki/api/v2/spaces"
Debug logs are written to: ~/.mcp/data/@aashari-mcp-server-atlassian-confluence.[session-id].log
Testing & Development
Using MCP Inspector
The MCP Inspector provides a visual interface for testing tools:
# Install the server globally
npm install -g @aashari/mcp-server-atlassian-confluence
# Run with MCP Inspector
npx @modelcontextprotocol/inspector node $(which mcp-atlassian-confluence)
Or use the built-in development command if you've cloned the repository:
npm run mcp:inspect
This starts the server in HTTP mode and opens the inspector UI in your browser.
HTTP Mode for Testing
You can run the server in HTTP mode to test with curl or other HTTP clients:
# Start server in HTTP mode
TRANSPORT_MODE=http npx -y @aashari/mcp-server-atlassian-confluence
The server will listen on http://localhost:3000/mcp by default. You can change the port:
Node.js/TypeScript MCP server for Atlassian Confluence. Provides tools enabling AI systems (LLMs) to list/get spaces & pages (content formatted as Markdown) and search via CQL. Connects AI seamlessly
Node.js/TypeScript MCP server for Atlassian Confluence. Provides tools enabling AI systems (LLMs) to list/get spaces & pages (content formatted as Markdown) and search via CQL. Connects AI seamlessly to Confluence knowledge bases using the standard MCP in
@aashari/mcp Server Atlassian Confluence 需要 API 金鑰嗎?
這筆記錄未列出必要的環境變數。
@aashari/mcp Server Atlassian Confluence 使用哪種傳輸方式?
@aashari/mcp Server Atlassian Confluence 透過 stdio 在本機執行。
@aashari/mcp Server Atlassian Confluence 上次檢查是什麼時候?
上次檢查時間:2026-11-17。 Node.js/TypeScript MCP server for Atlassian Confluence. Provides tools enabling AI systems (LLMs) to list/get spaces & pages (content formatted as Markdown) and search via CQL. Connects AI seamlessly to Confluence knowledge bases using the standard MCP in