SevenRooms MCP Server
Enables restaurant reservation management through SevenRooms API, allowing users to create reservations and query available time slots with guest details and party size information.
README
SevenRooms MCP Server
HTTP-based Model Context Protocol (MCP) server exposing tools and resources for SevenRooms reservations and availability. Uses StreamableHTTPServerTransport (HTTP + SSE) instead of stdio. Built with TypeScript and the official @modelcontextprotocol/sdk.
Features
make_reservationstool: Create restaurant reservations with guest detailsavailable_time_slotresource: Query available reservation time slots- TypeScript + Zod validation
- Streamable HTTP transport (JSON responses + optional SSE streaming)
- Ready for Azure App Service deployment via GitHub Actions
Project Structure
src/
server.ts # Main MCP HTTP server (/mcp endpoint)
tools/makeReservation.ts # make_reservations tool registration
resources/available_time_slot.ts # available_time_slot resource registration
test/ # Integration tests
.env.example # Example environment vars
tsconfig.json
package.json
.github/workflows/azure-deploy.yml # CI/CD workflow to Azure App Service
Environment Variables
| Name | Required | Description |
|---|---|---|
| SEVENROOMS_API_KEY | Yes | Authentication for SevenRooms API |
| SEVENROOMS_API_URL | Yes | Base API URL (e.g. https://api.sevenrooms.com) |
| PORT | Optional | Listening port (App Service sets automatically) |
Copy .env.example to .env for local use:
SEVENROOMS_API_KEY=your_key_here
SEVENROOMS_API_URL=https://api.sevenrooms.com
Local Development
npm install
npm run build
npm start # starts HTTP server on PORT (default 3000)
# or
npm run start:dev # tsx live-reload
Server endpoints:
POST /mcp– JSON-RPC MCP requests (initialize, tools/list, resources/list, tools/call, etc.)GET /mcp– SSE stream (if using streaming scenarios)DELETE /mcp– Close session (stateful mode; stateless here so optional)GET /health– Simple health probe
Sample Initialize Request
curl -X POST http://localhost:3000/mcp \
-H "Accept: application/json" \
-H "Content-Type: application/json" \
-d '{
"jsonrpc":"2.0",
"id":1,
"method":"initialize",
"params":{
"protocolVersion":"1.0",
"clientInfo":{"name":"local-client","version":"1.0"},
"capabilities":{}
}
}'
List Tools
curl -X POST http://localhost:3000/mcp \
-H "Accept: application/json" -H "Content-Type: application/json" \
-d '{"jsonrpc":"2.0","id":2,"method":"tools/list","params":{}}'
Call make_reservations Tool (example)
curl -X POST http://localhost:3000/mcp \
-H "Accept: application/json" -H "Content-Type: application/json" \
-d '{
"jsonrpc":"2.0",
"id":3,
"method":"tools/call",
"params":{
"name":"make_reservations",
"arguments":{
"date":"2025-12-20",
"time":"19:00",
"party_size":2,
"first_name":"Jane",
"last_name":"Doe",
"email":"jane@example.com",
"phone":"+15555550123"
}
}
}'
Tools & Resources Details
Tool: make_reservations
Input fields:
date, time, party_size, first_name, last_name, email, phone
Returns success or detailed error text from SevenRooms API.
Resource: available_time_slot
URI format:
available://YYYY-MM-DD/HH:MM/party_size
Returns JSON: { "available_times": ["18:00", "18:30", ...] } filtered for entries where type === 'book.
SevenRooms API Endpoints Used
- Reservations:
POST {SEVENROOMS_API_URL}/reservations - Availability:
GET {SEVENROOMS_API_URL}/availabilitywith query paramsdate,time,party_size
Adjust endpoint paths if your SevenRooms account differs.
Azure App Service Deployment
1. Create Azure Resources (CLI Example)
az group create -n sevenrooms-rg -l eastus
az appservice plan create -g sevenrooms-rg -n sevenrooms-plan --sku B1 --is-linux
az webapp create -g sevenrooms-rg -p sevenrooms-plan -n <WEBAPP_NAME> --runtime "NODE:18-lts"
2. Configure App Settings
az webapp config appsettings set -g sevenrooms-rg -n <WEBAPP_NAME> --settings \
SEVENROOMS_API_KEY="<secret>" \
SEVENROOMS_API_URL="https://api.sevenrooms.com"
Azure injects PORT automatically; do not hardcode it unless needed.
3. GitHub Secrets
Add in repo Settings → Actions → Secrets:
AZURE_WEBAPP_NAME= <WEBAPP_NAME>AZURE_WEBAPP_PUBLISH_PROFILE= publish profile XML pasted verbatim
4. Workflow (azure-deploy.yml)
Steps: checkout, setup Node + cache, install (skip if cache hit), test, build, list build output, deploy.
5. Verify Deployment
curl https://<WEBAPP_NAME>.azurewebsites.net/health
curl -X POST https://<WEBAPP_NAME>.azurewebsites.net/mcp \
-H "Accept: application/json" -H "Content-Type: application/json" \
-d '{"jsonrpc":"2.0","id":1,"method":"initialize","params":{"protocolVersion":"1.0","clientInfo":{"name":"remote","version":"1.0"},"capabilities":{}}}'
Logging & Observability
- Use
console.error()for server logs (stderr captured by App Service). - Add Application Insights: set
APPINSIGHTS_INSTRUMENTATIONKEYor connection string. - Consider structured logging (JSON) for easier analysis.
Troubleshooting
| Issue | Cause | Fix |
|---|---|---|
| 404 /mcp | Build missing or wrong start path | Ensure npm run build and start uses node build/server.js |
| 500 SevenRooms errors | Invalid API key or payload | Verify env vars & input fields |
| Timeout | Low SKU or network latency | Increase plan size / adjust retry |
| Not Acceptable | Missing Accept header | Use Accept: application/json (JSON-only enabled) |
| Capabilities error | Missing capabilities in initialize |
Include "capabilities":{} |
Production Recommendations
- Use at least B1 plan; scale out based on CPU or HTTP queue length.
- Restrict ingress with Access Restrictions or Front Door.
- Rotate
SEVENROOMS_API_KEYregularly. - Implement retries for transient SevenRooms failures (429, 5xx).
- Add rate limiting if exposed publicly.
Contributing
- Branch from
main - Implement changes in
src/ npm test && npm run build- PR and merge to trigger deployment
License
MIT (update as needed).
References
- MCP Docs: https://modelcontextprotocol.io/
- SDK Repo: https://github.com/modelcontextprotocol/typescript-sdk
- SevenRooms API Docs: (refer to account-specific portal)
SevenRooms MCP Server
A Model Context Protocol (MCP) server for managing restaurant reservations with SevenRooms. Built according to the MCP specification using TypeScript and the official @modelcontextprotocol/sdk.
Features
- make_reservations Tool: Create restaurant reservations with guest details
- available_time_slot Resource: Query available reservation time slots
- Fully typed with TypeScript and Zod validation
- STDIO-based communication for seamless MCP integration
- Azure App Service deployment ready
Project Structure
src/
index.ts # Main MCP server with tools and resources
test/
makeReservation.test.js # Unit tests for make_reservations tool
availableTimeSlot.test.js # Unit tests for available_time_slot resource
.env.example # Example environment variables
tsconfig.json # TypeScript configuration
package.json # Project dependencies and scripts
.github/
workflows/
azure-deploy.yml # GitHub Actions CI/CD workflow
Prerequisites
- Node.js >= 16
- npm
- SevenRooms API key and base URL
Local Development
1. Install Dependencies
npm install
2. Configure Environment Variables
Copy .env.example to .env and fill in your SevenRooms credentials:
cp .env.example .env
Then edit .env:
SEVENROOMS_API_KEY=your_sevenrooms_api_key_here
SEVENROOMS_API_URL=https://api.sevenrooms.com
3. Build the Server
npm run build
This compiles TypeScript to JavaScript in the build/ directory.
4. Run Tests
npm test
Expected output:
available_time_slot resource
✓ throws when required fields missing
✓ returns only times with type == "book"
make_reservations tool
✓ throws when required fields missing
✓ calls SevenRooms reservations endpoint and returns data
4 passing
5. Start the Server
For production:
npm start
For development with auto-reload:
npm run start:dev
The server will start on stdio, ready for MCP clients to connect.
MCP Tools & Resources
Tool: make_reservations
Makes a reservation at a restaurant through SevenRooms.
Input Schema:
{
"date": "YYYY-MM-DD (e.g., 2025-12-25)",
"time": "HH:MM (e.g., 19:00)",
"party_size": 4,
"first_name": "John",
"last_name": "Doe",
"email": "john@example.com",
"phone": "555-1234"
}
Response: Returns the SevenRooms API response with reservation confirmation details (ID, status, etc.).
Resource: available_time_slot
Queries available reservation time slots for a given date, time, and party size.
URI Format:
available://YYYY-MM-DD/HH:MM/party_size
Example:
available://2025-12-25/19:00/4
Response:
{
"available_times": ["18:00", "18:30", "19:00", "19:30"]
}
The resource automatically filters to only include times where type === 'book' from the SevenRooms availability API.
SevenRooms API Integration
The server calls the following SevenRooms endpoints:
Reservations Endpoint
- URL:
{SEVENROOMS_API_URL}/reservations - Method: POST
- Auth: Bearer token in
Authorizationheader - Payload:
{ "datetime": "2025-12-25T19:00", "party_size": 4, "guest": { "first_name": "John", "last_name": "Doe", "email": "john@example.com", "phone": "555-1234" } }
Availability Endpoint
- URL:
{SEVENROOMS_API_URL}/availability - Method: GET
- Auth: Bearer token in
Authorizationheader - Query Params:
date,time,party_size
Note: Adjust endpoints and request payloads if your SevenRooms API account requires different paths or authentication methods.
Testing
The project includes unit tests using Mocha, Chai, and Nock (for HTTP mocking).
Run tests:
npm test
Tests cover:
- Input validation for both tool and resource
- Mocked SevenRooms API responses
- Filtering logic for available time slots (type === 'book')
Azure Deployment
Prerequisites
- Azure App Service instance
- GitHub repository with this code
- Azure publish profile exported
Setup Steps
-
Create Azure App Service
- Create a new Web App (Node.js 18 LTS or later)
- Copy the publish profile XML
-
Configure GitHub Secrets
- In your GitHub repository, go to Settings → Secrets and variables → Actions
- Add these secrets:
AZURE_WEBAPP_NAME— The name of your Azure Web AppAZURE_WEBAPP_PUBLISH_PROFILE— The publish profile XML content
-
Set App Service Configuration
- In Azure portal, go to your App Service → Configuration
- Add these application settings:
SEVENROOMS_API_KEY— Your SevenRooms API keySEVENROOMS_API_URL— Your SevenRooms API base URL (e.g.,https://api.sevenrooms.com)PORT— Leave blank to auto-bind (Azure sets this automatically)
-
Deploy
- Push to
mainbranch - GitHub Actions workflow runs automatically
- Deployment proceeds to Azure App Service
- Push to
The workflow is defined in .github/workflows/azure-deploy.yml and:
- Checks out the code
- Installs dependencies
- Builds the TypeScript
- Deploys using the publish profile
Verify Deployment
After deployment, you can:
- Check the App Service activity log in Azure portal
- Review GitHub Actions workflow run logs on GitHub
- Test the server by connecting via MCP client pointing to the deployed instance
Logging
Important: The MCP server uses stdio for communication, so logging is restricted:
- ✅ Use
console.error()for logging (writes to stderr, safe for STDIO-based MCP) - ❌ Never use
console.log()(writes to stdout, corrupts MCP JSON-RPC messages)
For production, consider:
- Redirecting logs to Application Insights via Azure SDK
- Using structured logging libraries that write to stderr
- Checking Azure App Service logs in the Azure portal
Troubleshooting
"Cannot find module '@modelcontextprotocol/sdk'"
- Run
npm installto ensure all dependencies are installed - Check that
@modelcontextprotocol/sdkis inpackage.jsondependencies
Build errors (TypeScript)
- Ensure Node.js >= 16 is installed
- Delete
build/andnode_modules/and reinstall:npm install && npm run build
Tests failing
- Ensure
.envfile exists (even if empty) or set env vars in your shell before running tests - Check that nock mocks match your actual SevenRooms API requests
Server not connecting as MCP client
- Ensure the server runs without throwing errors:
npm start(should print to stderr) - Verify the MCP client is correctly configured to call the server
Contributing
- Make changes to
src/index.ts - Run tests:
npm test - Build:
npm run build - Commit and push to trigger CI/CD
License
MIT
References
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。