RefMD MCP Server
Enables chatbots to browse, search, create, and edit Markdown documents stored in RefMD through a hosted SSE endpoint with OAuth 2.1 authentication.
README
RefMD MCP Server
A Model Context Protocol server that exposes RefMD documents over a hosted SSE endpoint, so chatbots can browse and edit Markdown through RefMD's API.
Features
- Streamable HTTP/SSE transport compatible with hosted MCP clients (e.g. Claude, Cursor)
- Resource template
refmd://document/{id}to read document Markdown and metadata - Tools for listing, searching, creating, and updating documents via the RefMD API
Prerequisites
- Node.js 18+
- Access to an existing RefMD instance
- Either a personal access token (JWT) or credentials that can log in via
/api/auth/login
Configuration
The server now authenticates users via OAuth 2.1 with PKCE. Configure it with the following variables:
| Variable | Description |
|---|---|
REFMD_API_BASE |
Required. Base URL of your RefMD API (e.g. https://refmd.example.com). |
OAUTH_CLIENT_IDS |
Comma-separated list of allowed client_id values. Leave empty to allow any. |
OAUTH_ALLOWED_REDIRECTS |
Comma-separated list of allowed redirect URIs. Defaults to allowing HTTPS URLs and http://localhost. Include ChatGPT’s callback URL (https://chat.openai.com/aip/mcp/oauth/callback). |
OAUTH_ISSUER |
Optional public issuer URL (defaults to the current request origin). |
ACCESS_TOKEN_TTL_SECONDS |
Optional access-token lifetime (default 3600). |
REFRESH_TOKEN_TTL_SECONDS |
Optional refresh-token lifetime (default 2592000, i.e. 30 days). |
MCP_DB_DRIVER |
Optional. Set to sqlite, postgres, or mysql to persist OAuth tokens. Defaults to in-memory storage. |
MCP_DB_URL |
Optional connection string for the configured driver (e.g. postgres://user:pass@host/db). For SQLite you may omit this and use MCP_DB_SQLITE_PATH. |
MCP_DB_SQLITE_PATH |
Optional filesystem path for SQLite storage (defaults to ./data/refmd-mcp.sqlite). Accepts plain paths or file:/// URLs; ensure the path resolves to persistent storage when using SQLite. |
PORT / HOST |
Optional listen port / host (defaults: 3334 / 0.0.0.0). |
ℹ️ Install the appropriate database driver when enabling persistence:
npm install better-sqlite3for SQLite,npm install pgfor PostgreSQL, ornpm install mysql2for MySQL/MariaDB.
Install & Build
cd mcp-server
npm install
npm run build
Run
npm start
REFMD_API_BASE="https://refmd.example.com" \
OAUTH_CLIENT_IDS="chatgpt-connector" \
OAUTH_ALLOWED_REDIRECTS="https://chat.openai.com/aip/mcp/oauth/callback" \
npm start
The server exposes two transports:
http://<host>:<port>/sse— SSE transport (compatible with Claude SSE etc.)http://<host>:<port>/mcp— Streamable HTTP transport (one-shot POST per exchange)
OAuth flow
- Configure your client (e.g. ChatGPT custom connector) with:
- Authorization URL:
https://your-domain.example.com/oauth/authorize - Token URL:
https://your-domain.example.com/oauth/token - Revocation URL:
https://your-domain.example.com/oauth/revoke - Scopes: (leave blank)
- PKCE: enabled (ChatGPT uses S256 automatically)
- Authorization URL:
- When prompted, the browser shows the RefMD MCP consent page. Paste a RefMD API token generated from Profile → API tokens and approve.
- The connector receives an access token and can call
/sseor/mcpwithAuthorization: Bearer <token>.
Tokens can be revoked from RefMD (profile page) or via POST /oauth/revoke.
Run with Docker
# Build image
docker build -t refmd-mcp .
docker run -p 3334:3334 \
-e REFMD_API_BASE="https://refmd.example.com" \
-e OAUTH_CLIENT_IDS="chatgpt-connector" \
-e OAUTH_ALLOWED_REDIRECTS="https://chat.openai.com/aip/mcp/oauth/callback" \
-e MCP_DB_DRIVER="sqlite" \
-e MCP_DB_SQLITE_PATH="/data/refmd-mcp.sqlite" \
-v refmd-mcp-data:/data \
refmd-mcp
Mount a persistent volume (as shown above) so the SQLite database file survives container restarts.
Connecting a Chat Client
- Claude (CLI):
claude mcp add --transport sse refmd https://your-domain.example.com/sse - Cursor / VS Code / MCP Inspector: choose an SSE transport and supply the same URL.
Once connected, resources appear under refmd://document/{id}. Available tools include refmd-list-documents, refmd-search-documents, refmd-create-document, refmd-read-document, refmd-update-document-content, and more.
Release workflow
The GitHub Actions workflow CI MCP Server ships the container image. It runs automatically on pushes/PRs touching mcp-server and publishes to GHCR when:
- the push is a tag matching
mcp-server-v*(versioned release), or - the workflow is manually triggered with
publish=true.
Tags published to ghcr.io/<owner>/refmd-mcp include semantic versions (1.2.0, 1.2, 1), the raw git tag, and latest. Use the extra-tag input for additional labels when invoking the workflow manually.
Development
Run in watch mode with TSX:
npm run dev
Any code changes require a rebuild (npm run build) before deploying or running with npm start.
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。