Oura MCP Server
Enables Claude and other MCP clients to access Oura Ring data, including daily activity, sleep, readiness, stress, and more, via a remote Streamable HTTP server.
README
Oura MCP Server
A remote Model Context Protocol (MCP) server for accessing Oura Ring data over Streamable HTTP.
This server is designed for deployment behind HTTPS, such as a self-hosted Coolify instance, so it can be added to Claude.ai as a Custom Connector. It uses the MCP TypeScript SDK's StreamableHTTPServerTransport, not stdio and not the legacy SSE transport.
Setup
Prerequisites
- Node.js 22+
- pnpm
- Oura account
Installation
- Clone the repository
- Run:
pnpm install
pnpm run build
Configuration
Environment Variables
Create a .env file for local development only:
PORT=3000
MCP_ALLOWED_HOSTS=oura-mcp.mydomain.com
OURA_PERSONAL_ACCESS_TOKEN=your_token
Do not commit .env or any real Oura token. In Coolify, set OURA_PERSONAL_ACCESS_TOKEN in the Resource environment variable UI.
You can also set MCP_ALLOWED_HOSTS to your connector hostname, for example oura-mcp.mydomain.com, to restrict accepted Host headers.
Obtaining An Oura Token
- Log in to Oura Cloud Console.
- Create a Personal Access Token at Personal Access Tokens.
- Store it only as
OURA_PERSONAL_ACCESS_TOKENin your deployment environment.
Local Usage
pnpm run build
PORT=3000 OURA_PERSONAL_ACCESS_TOKEN=placeholder pnpm start
The MCP endpoint is:
http://localhost:3000/mcp
Docker
Build and run locally:
docker build -t oura-mcp .
docker run --rm -p 3000:3000 -e OURA_PERSONAL_ACCESS_TOKEN=your_token oura-mcp
Coolify Deployment
- Push this repository to your GitHub account, or use Coolify's Public Repository option with your fork URL.
- In Coolify, create a new Project if needed.
- Add a new Resource.
- Choose your Git repository, or choose Public Repository and paste the repository URL.
- Select Dockerfile as the build pack.
- Confirm the Dockerfile path is
Dockerfile. - Set Port Exposes to
3000. - In Environment Variables, add:
Do not put the token in code,OURA_PERSONAL_ACCESS_TOKEN=<your Oura personal access token>.env.example, build arguments, or logs. - Optional but recommended: add
MCP_ALLOWED_HOSTSwith only the hostname from your FQDN, for example:MCP_ALLOWED_HOSTS=oura-mcp.mydomain.com - Set the Resource FQDN/domain field, for example:
https://oura-mcp.mydomain.com
- Deploy. Coolify will route HTTPS traffic to the container and provision a Let's Encrypt certificate for the FQDN.
Your final Claude Custom Connector URL should be:
https://oura-mcp.mydomain.com/mcp
MCP Handshake Check
After deployment, replace the URL below with your actual HTTPS FQDN:
curl -i https://oura-mcp.mydomain.com/mcp \
-H 'Content-Type: application/json' \
-H 'Accept: application/json, text/event-stream' \
-d '{"jsonrpc":"2.0","id":1,"method":"initialize","params":{"protocolVersion":"2025-06-18","capabilities":{},"clientInfo":{"name":"curl-handshake","version":"1.0.0"}}}'
A healthy response returns HTTP 200, an mcp-session-id response header, and a JSON-RPC result containing the server info for oura-provider.
Available Resources
personal_info- User profiledaily_activity- Activity summariesdaily_readiness- Readiness scoresdaily_sleep- Sleep summariessleep- Detailed sleep datasleep_time- Sleep timingworkout- Workout datasession- Session datadaily_spo2- SpO2 measurementsrest_mode_period- Rest periodsring_configuration- Ring configdaily_stress- Stress metricsdaily_resilience- Resilience metricsdaily_cardiovascular_age- CV agevO2_max- VO2 max data
Available Tools
For date-based resources, use tools like get_daily_sleep with startDate and endDate parameters (YYYY-MM-DD).
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。