Calorie Tracker MCP Server
Tracks daily calorie intake with accurate BMR/TDEE calculations, food logging, and profile management via Cloudflare Workers and D1 database.
README
Calorie Tracker MCP Server
MCP server for tracking daily calorie intake with accurate BMR/TDEE calculations. Built on Cloudflare Workers with D1 database.
Features
- Food Tracking: Add, update, delete food entries with macros
- Profile Management: BMR/TDEE calculations using Mifflin-St Jeor equation
- Historical Data: Track weight, body composition over time
- Secure: OAuth 2.1 authentication with PKCE support
Quick Setup
pnpm install
# Create resources and update wrangler.jsonc with IDs
npx wrangler d1 create calorie-tracker
npx wrangler kv:namespace create OAUTH_KV
# Run migrations and start dev server
npx wrangler d1 migrations apply calorie-tracker --local
pnpm run dev
OAuth Client Registration
Set an admin API key:
# Development: Update wrangler.jsonc vars.ADMIN_API_KEY
# Production: Use secrets
npx wrangler secret put ADMIN_API_KEY
Register an OAuth client to obtain credentials:
curl -X POST http://localhost:8787/oauth/register \
-H "Content-Type: application/json" \
-H "X-API-Key: YOUR_ADMIN_API_KEY" \
-d '{
"client_name": "Claude Desktop",
"redirect_uris": ["http://127.0.0.1/callback", "http://localhost/callback"],
"user_id": "admin"
}'
Note: Claude Desktop will specify its callback URL during the OAuth flow. If you get a redirect URI mismatch error, check Claude Desktop's OAuth logs for the actual callback URL it expects.
Response:
{
"client_id": "uuid-client-id",
"client_secret": "generated-secret",
"client_name": "Claude Desktop"
}
Claude Configuration
Option 1: Bearer Token (Personal/Pro/Max)
Simple authentication using client secret as bearer token:
-
Register OAuth client:
curl -X POST http://localhost:8787/oauth/register \ -H "Content-Type: application/json" \ -H "X-API-Key: YOUR_ADMIN_API_KEY" \ -d '{ "client_name": "Claude Desktop", "redirect_uris": ["https://claude.ai/api/mcp/auth_callback"], "user_id": "admin" }'Save the
client_secretfrom response. -
Claude Desktop config (
claude_desktop_config.json):{ "mcpServers": { "calorie-tracker": { "command": "npx", "args": ["mcp-remote", "http://localhost:8787/sse"], "env": { "BEARER_TOKEN": "<client_secret from registration>" } } } }
Option 2: OAuth with Client Credentials (Team/Enterprise Only)
Private connector with Client ID/Secret:
-
Register OAuth client with callback URL
https://claude.ai/api/mcp/auth_callback -
Add custom connector in Claude:
- Name: Calorie Tracker
- Remote MCP server URL:
https://YOUR-WORKER.workers.dev/sse - OAuth Client ID:
<client_id from registration> - OAuth Client Secret:
<client_secret from registration>
Production Deployment
Deploy to Cloudflare Workers for mobile/web access:
pnpm run deploy
pnpm run migrate
npx wrangler secret put ADMIN_API_KEY
Tools
Food Tracking:
list_entries- List food entries with paginationadd_entry- Add food entry with macrosupdate_entry- Update existing entrydelete_entry- Delete entry
Profile Management:
get_profile- Get profile with BMR/TDEE calculationsupdate_profile- Update profile data (height, weight, activity level)get_profile_history- Historical tracking data
Admin:
register_user- Register new userrevoke_user- Revoke user access
BMR/TDEE Calculations
Uses Mifflin-St Jeor equation (gold standard):
Male: BMR = (10 × weight) + (6.25 × height) - (5 × age) + 5
Female: BMR = (10 × weight) + (6.25 × height) - (5 × age) - 161
TDEE: BMR × Activity Multiplier (1.2 - 1.9)
Development
pnpm test # Run tests
pnpm run type-check # TypeScript validation
pnpm run deploy # Deploy to production
pnpm run migrate # Run migrations on remote
Tech Stack
- Runtime: Cloudflare Workers
- Database: D1 (SQLite)
- Language: TypeScript
- Validation: Zod
- Testing: Vitest
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。