AI Interface for Outline Knowledge Systems (mcp-outline)

This repository details a Model Context Protocol (MCP) intermediary service engineered to bridge large language models (like Claude) with the Outline document hosting platform (https://www.getoutline.com).

Core Capabilities

This implementation offers robust connectivity to Outline's document ecosystem, including:

• Content Discovery: Execute keyword-based searches across all stored documents.
• Structure Mapping: Enumerate organizational structures (collections) and their hierarchical document layouts.
• Data Retrieval: Fetch and export the full content of specified documents, typically formatted as Markdown.
• Feedback Loop: Review existing annotations and introduce new commentary on specific entries.
• Creation Workflow: Generate novel documents within designated organizational units.
• Maintenance Tasks: Modify existing document text or relocate items between different collections.
• Dependency Mapping: Identify and list documents that reference a specific target document (backlinks).

Deployment via Docker (Recommended)

Leveraging Docker simplifies dependency management for this Python-based MCP server.

1. Ensure Docker is installed and operational.
2. Build the necessary container image: docker buildx build -t mcp-outline .
3. Configure the server connection within your AI client (e.g., Cursor):

{ "mcpServers": { "outline-doc-agent": { "command": "docker", "args": [ "run", "-i", "--rm", "--init", "-e", "DOCKER_CONTAINER=true", "-e", "OUTLINE_API_KEY", "-e", "OUTLINE_API_URL", "mcp-outline" ], "env": { "OUTLINE_API_KEY": "", "OUTLINE_API_URL": "" } } } }

Note: OUTLINE_API_URL defaults to the standard cloud endpoint if omitted. 4. For debugging connectivity, utilize the MCP inspector utility: bash npx @modelcontextprotocol/inspector docker run -i --rm --init -e DOCKER_CONTAINER=true --env-file .env mcp-outline

Development Setup

Prerequisites

• Python runtime version 3.10 or later.
• Active Outline account with API key permissions.

Installation Steps

Clone the source repository and install dependencies in editable mode:

bash git clone https://github.com/Vortiago/mcp-outline.git cd mcp-outline uv pip install -e " சகdev]"

Configuration File (.env)

Set up your authentication credentials in a .env file:

Outline Authentication Token

OUTLINE_API_KEY=your_outline_api_key_here

Optional: Specify URL for self-hosted instances

OUTLINE_API_URL=https://your-custom-domain.com/api

Running the Service

Start the server for development, optionally integrating with the inspector tool:

bash

Standard development launch

mcp dev src/mcp_outline/server.py

Or use the provided launch script

./start_server.sh

Interaction Examples

Querying:

Locate all articles pertaining to "Q3 financial projections".

Structure Review:

Display the contents of the 'Engineering Standards' knowledge base.

Content Retrieval:

Retrieve and format the data from the entry identified by ID "docId123".

Annotation Activity:

Append the feedback "Please verify all source citations" to document "docId123".

Project Metadata

Licensed under the MIT License. Acknowledgements include the [MCP Python SDK] and the underlying [Outline REST API].