Outline Wiki MCP Server

English | | 日本語 | 中文

A Model Context Protocol (MCP) server that enables LLMs to interact with Outline wiki through structured API calls. This server provides document management, search, collections, comments, and AI-powered smart features including RAG-based Q&A.

Why This Server?

Most Outline MCP servers provide basic API wrappers. This one adds optional Smart Features:

Feature	What it does
ask_wiki	Ask questions in natural language, get answers based on your wiki content (RAG)
find_related	Find semantically similar documents, not just keyword matches
summarize_document	Generate summaries of long documents
suggest_tags	Get tag suggestions based on content analysis

When you might need this:

• Your team's wiki has grown large and search isn't enough
• You want to query your documentation conversationally
• You need semantic search across your knowledge base

When basic MCP is sufficient:

• You only need CRUD operations on documents
• You don't want to set up OpenAI API
• Your wiki is small and well-organized

Smart features require ENABLE_SMART_FEATURES=true and an OpenAI API key. Without these, the server works as a standard Outline MCP.

Example Usage

User: "What's our policy on remote work?"
 ask_wiki searches your wiki and returns an answer with source links

User: "Find documents related to the onboarding guide"
 find_related returns semantically similar docs (not just keyword matches)

User: "Summarize the Q4 planning document"
 summarize_document generates a concise summary in your preferred language

Supported Clients

Client	Tools	Resources	Prompts
Claude Desktop
Claude Code
VS Code GitHub Copilot
Cursor		-
Windsurf		-	-
ChatGPT Desktop		-	-

Getting Started

Requirements

• Node.js 18.0.0 or higher
• Outline instance with API access
• (Optional) OpenAI API key for smart features

Getting Your Outline API Token

1. Log in to your Outline instance
2. Go to Settings API
3. Click Create API Key
4. Copy the generated token (starts with ol_api_)

Installation

Add to your Claude Desktop configuration:

• macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
• Windows: %APPDATA%\Claude\claude_desktop_config.json
• Linux: ~/.config/Claude/claude_desktop_config.json

{
  "mcpServers": {
    "outline": {
      "command": "npx",
      "args": ["-y", "outline-smart-mcp"],
      "env": {
        "OUTLINE_URL": "https://your-outline-instance.com",
        "OUTLINE_API_TOKEN": "ol_api_xxxxxxxxxxxxx"
      }
    }
  }
}

Run the following command:

claude mcp add outline -e OUTLINE_URL=https://your-outline-instance.com -e OUTLINE_API_TOKEN=ol_api_xxxxxxxxxxxxx -- npx -y outline-smart-mcp

Or add to ~/.claude.json (global) or .mcp.json (project-local):

{
  "mcpServers": {
    "outline": {
      "command": "npx",
      "args": ["-y", "outline-smart-mcp"],
      "env": {
        "OUTLINE_URL": "https://your-outline-instance.com",
        "OUTLINE_API_TOKEN": "ol_api_xxxxxxxxxxxxx"
      }
    }
  }
}

Note: The ~/.claude/settings.json file is ignored for MCP servers. Use ~/.claude.json or .mcp.json instead.

Add to your VS Code settings (.vscode/mcp.json):

{
  "servers": {
    "outline": {
      "command": "npx",
      "args": ["-y", "outline-smart-mcp"],
      "env": {
        "OUTLINE_URL": "https://your-outline-instance.com",
        "OUTLINE_API_TOKEN": "ol_api_xxxxxxxxxxxxx"
      }
    }
  }
}

Add to Cursor MCP settings (~/.cursor/mcp.json):

{
  "mcpServers": {
    "outline": {
      "command": "npx",
      "args": ["-y", "outline-smart-mcp"],
      "env": {
        "OUTLINE_URL": "https://your-outline-instance.com",
        "OUTLINE_API_TOKEN": "ol_api_xxxxxxxxxxxxx"
      }
    }
  }
}

Add to Windsurf MCP settings (~/.codeium/windsurf/mcp_config.json):

{
  "mcpServers": {
    "outline": {
      "command": "npx",
      "args": ["-y", "outline-smart-mcp"],
      "env": {
        "OUTLINE_URL": "https://your-outline-instance.com",
        "OUTLINE_API_TOKEN": "ol_api_xxxxxxxxxxxxx"
      }
    }
  }
}

ChatGPT supports MCP through its desktop app. Add the server in Settings MCP Servers with:

• Command: npx
• Arguments: -y outline-smart-mcp
• Environment variables as shown above

Configuration

Environment Variables

Variable	Description	Required	Default
OUTLINE_URL	Your Outline instance URL	Yes	https://app.getoutline.com
OUTLINE_API_TOKEN	Your Outline API token	Yes	-
READ_ONLY	Enable read-only mode	No	false
DISABLE_DELETE	Disable delete operations	No	false
MAX_RETRIES	API retry attempts	No	3
RETRY_DELAY_MS	Retry delay (ms)	No	1000
ENABLE_SMART_FEATURES	Enable AI features	No	false
OPENAI_API_KEY	OpenAI API key	No*	-

* Required when ENABLE_SMART_FEATURES=true

Smart Features Configuration

To enable AI-powered features (RAG Q&A, summarization, etc.), add these to your config:

{
  "mcpServers": {
    "outline": {
      "command": "npx",
      "args": ["-y", "outline-smart-mcp"],
      "env": {
        "OUTLINE_URL": "https://your-outline-instance.com",
        "OUTLINE_API_TOKEN": "ol_api_xxxxxxxxxxxxx",
        "ENABLE_SMART_FEATURES": "true",
        "OPENAI_API_KEY": "sk-xxxxxxxxxxxxx"
      }
    }
  }
}

Tools

Search & Discovery

Tool	Description
search_documents	Search documents by keyword with pagination
get_document_id_from_title	Find document ID by title
list_collections	Get all collections
get_collection_structure	Get document hierarchy in a collection
list_recent_documents	Get recently modified documents

Document Operations

Tool	Description
get_document	Get full document content by ID
export_document	Export document in Markdown
create_document	Create a new document
update_document	Update document (supports append)
move_document	Move document to another location

Document Lifecycle

Tool	Description
archive_document	Archive a document
unarchive_document	Restore archived document
delete_document	Delete document (soft/permanent)
restore_document	Restore from trash
list_archived_documents	List archived documents
list_trash	List trashed documents

Comments & Collaboration

Tool	Description
add_comment	Add comment (supports replies)
list_document_comments	Get document comments
get_comment	Get specific comment
get_document_backlinks	Find linking documents

Collection Management

Tool	Description
create_collection	Create collection
update_collection	Update collection
delete_collection	Delete collection
export_collection	Export collection
export_all_collections	Export all collections

Batch Operations

Tool	Description
batch_create_documents	Create multiple documents
batch_update_documents	Update multiple documents
batch_move_documents	Move multiple documents
batch_archive_documents	Archive multiple documents
batch_delete_documents	Delete multiple documents

Smart Features (AI-Powered)

Requires ENABLE_SMART_FEATURES=true and OPENAI_API_KEY.

Tool	Description
smart_status	Check status and indexed count
sync_knowledge	Sync docs to vector database
ask_wiki	RAG-based Q&A on wiki content
summarize_document	Generate AI summary
suggest_tags	AI-suggested tags
find_related	Find semantically related docs
generate_diagram	Generate Mermaid diagrams

Smart Features Usage

# 1. First, sync your wiki documents
sync_knowledge

# 2. Ask questions about your wiki
ask_wiki: "What is our deployment process?"

# 3. Summarize long documents
summarize_document: { documentId: "doc-id", language: "Korean" }

# 4. Find related content
find_related: { documentId: "doc-id", limit: 5 }

Technology Stack

Component	Technology
Vector Database	LanceDB (embedded)
Embeddings	OpenAI text-embedding-3-small
LLM	GPT-4o-mini
Text Chunking	LangChain

Safety Features

Read-Only Mode

READ_ONLY=true

Restricts to read operations only: search, get, export, list operations, and all smart features.

Disable Delete

DISABLE_DELETE=true

Blocks: delete_document, delete_collection, batch_delete_documents

Development

# Clone repository
git clone https://github.com/huiseo/outline-wiki-mcp.git
cd outline-wiki-mcp

# Install dependencies
npm install

# Build
npm run build

# Run tests
npm test

# Type check
npm run typecheck

License

MIT License - see LICENSE for details.

Links

• Outline Wiki
• Outline API Docs
• Model Context Protocol
• MCP Clients