Outline MCP Server

A Model Context Protocol (MCP) server that provides AI tools with full access to Outline — an open-source wiki and knowledge base.

Built with TypeScript, deployed via Docker MCP Gateway. Works with any MCP-compatible AI tool: Claude Code, Gemini CLI, Codex, and more.

Tools

| Tool | Description | |------|-------------| | search_documents | Full-text search across all workspace documents | | get_document | Read full document content (markdown) by ID | | create_document | Create a new document in a collection | | update_document | Update title or content, with append mode support | | delete_document | Move to trash (30-day recovery) or permanently delete | | archive_document | Archive an outdated document | | list_collections | List all accessible collections | | get_collection | Get collection details | | list_documents | List documents, optionally filtered by collection | | get_collection_structure | Get hierarchical document tree of a collection |

Prerequisites

• Docker Desktop with MCP Toolkit enabled
• An Outline instance with an API token (Settings > API)

Quick Start

1. Clone and build

git clone https://github.com/gupila/mcp-server-outline.git
cd mcp-server-outline
npm install
npm run build
docker build -t outline-mcp-server .

2. Configure secrets

docker mcp secret set OUTLINE_API_TOKEN="ol_api_your_token_here"
docker mcp secret set OUTLINE_BASE_URL="https://your-outline-instance.com"

3. Register with Docker MCP Gateway

Add to ~/.docker/mcp/catalogs/my-servers.yaml:

  outline:
    description: "Outline MCP Server"
    title: Outline
    type: server
    image: outline-mcp-server:latest
    ref: ""
    tools:
      - name: search_documents
      - name: get_document
      - name: create_document
      - name: update_document
      - name: delete_document
      - name: archive_document
      - name: list_collections
      - name: get_collection
      - name: list_documents
      - name: get_collection_structure
    secrets:
      - name: OUTLINE_API_TOKEN
        env: OUTLINE_API_TOKEN
      - name: OUTLINE_BASE_URL
        env: OUTLINE_BASE_URL
    prompts: 0
    resources: {}

Add to ~/.docker/mcp/registry.yaml:

  outline:
    ref: ""

4. Verify

docker mcp server list | grep outline

Usage

Once connected, ask your AI tool naturally:

• "Search Outline for onboarding docs"
• "List all collections"
• "Read the API guidelines document"
• "Create a new document called 'Sprint Retrospective' in the Engineering collection"
• "Append the deployment steps to the runbook"

Architecture

AI Tool (Claude Code / Gemini / Codex)
        │
        ▼
Docker MCP Gateway
        │
        ▼
Outline MCP Server (this project)
        │
        ▼
Outline REST API (/api/documents.*, /api/collections.*)

• Single-file TypeScript server (src/index.ts)
• Native fetch — no external HTTP dependencies
• Bearer token auth via Docker secrets
• 15s timeout on all API calls
• Non-root Docker container for security

Development

npm install          # Install dependencies
npm run typecheck    # Type check
npm run build        # Compile TypeScript
npm run dev          # Watch mode (auto-reload)
npm start            # Run compiled server

Adding a new tool

1. Add the tool function returning Promise<string> in src/index.ts
2. Add the tool definition to the TOOLS array
3. Add the case to the CallToolRequestSchema handler switch
4. Rebuild: npm run build && docker build -t outline-mcp-server .
5. Update your catalog YAML with the new tool name

License

MIT