Outline MCP Server
Enables AI agents to manage documents and collections in Outline wiki through natural language commands.
README
Outline MCP Server
Overview
This is a Model Context Protocol (MCP) server implementation that integrates with Outline. The server provides tools for AI agents (mostly Cursor) to interact with Outline's API, enabling document and collection management operations.
Features
- Search documents by query terms
- Create, retrieve, update, and delete documents
- List documents within collections
- Move documents between collections
- Manage collections (create, update, delete, list)
Installation
Prerequisites
- Node.js (v16 or later)
- npm or yarn
- Outline API key
Setup
- Clone the repository:
git clone https://github.com/fellowapp/mcp-outline.git
cd mcp-outline
- Install dependencies:
npm install
-
Create a
.envfile based on the.env.templatefile -
Get you Outline API token from your personal settings.
-
Connect to your MCP client (like Cursor) NOTE: if you are connecting to Cursor, you technically don't need to do the whole
.envfile setup, but it can be usefull for local testing. -
For connecting to Cursor, add this server to either your global or local project MCP settings with this:
"mcp-outline": {
"command": "node",
"args": [
"<FULL_PATH_OF_CLONED_OUTLINE_MCP_REPO>/src/index.js"
],
"env": {
"API_URL": "https://dev.fellow.wiki/api",
"API_KEY": "<OUTLINE_API_KEY>"
}
}
Architecture
This project implements the Model Context Protocol (MCP) standard for tool-based interactions. The architecture consists of:
src/
├── index.js # Main server entry point
├── outline.js # Outline API client configuration
├── tools/
│ ├── handlers.js # Centralized tool handlers mapping
│ ├── toolSchemas.js # Centralized tool schemas
│ ├── document/ # Document-related tools
│ │ ├── index.js # Export all document tools
│ │ ├── create.js # Create document tool
│ │ ├── delete.js # Delete document tool
│ │ ├── get.js # Get document tool
│ │ ├── list.js # List documents tool
│ │ ├── move.js # Move document tool
│ │ ├── search.js # Search documents tool
│ │ └── update.js # Update document tool
The server handles incoming requests through a stdio transport layer and routes them to the appropriate tool handler based on the requested tool name.
Development
Adding New Tools
To add a new tool:
- Create a new file in the appropriate directory (e.g.,
src/tools/document/my-tool.js) - Follow this template:
import { ErrorCode, McpError } from "@modelcontextprotocol/sdk/types.js";
import { outlineClient } from "../../outline.js";
const toolSchema = {
name: "my_tool_name",
description: "Description of what this tool does",
inputSchema: {
type: "object",
properties: {
// Define your parameters here
param1: {
type: "string",
description: "Description of param1",
},
},
required: ["param1"],
},
};
async function myToolHandler({ param1 }) {
try {
// Implement your tool logic here
const response = await outlineClient.post("/endpoint.name", {
params: {
param1,
},
});
return response.data.data;
} catch (error) {
console.error(error);
throw new McpError(ErrorCode.InvalidRequest, error.message);
}
}
export { toolSchema, myToolHandler as handler };
- Add your tool to the appropriate
index.jsfile
Credits
This project uses:
- Model Context Protocol TypeScript SDK
- Outline API
- Axios for HTTP requests
This README was (mostly) generated using Cursor.
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。