beaver-mcp
Provides Google Maps search, navigation, and travel time estimation alongside Lunch Money financial management, all accessible via a Cloudflare Workers MCP server with OAuth authentication.
README
beaver-mcp
Personal remote MCP server on Cloudflare Workers, connected to Claude as a custom connector at https://mcp.alanko.dev/mcp.
Tools
| Tool | What it does |
|---|---|
lunchmoney_list_transactions |
List/filter transactions (date range, category, tag, status, pending) |
lunchmoney_get_transaction |
Single transaction with split/group children |
lunchmoney_update_transaction |
Update payee, category, notes, status, tags, amount |
lunchmoney_create_transaction |
Create a manual transaction |
lunchmoney_list_categories / _tags / _accounts |
ID lookups for updates |
Auth: the endpoint is public, so /mcp sits behind OAuth (@cloudflare/workers-oauth-provider). Claude signs in via GitHub; only logins in the ALLOWED_GITHUB_LOGINS secret may authorize (empty = deny all).
The server is structured so any API becomes MCP tools with one folder (client.ts + tools.ts) and one line in src/index.ts — see the recipe in CLAUDE.md.
Setup
1. GitHub OAuth Apps (one-time)
GitHub → Settings → Developer settings → OAuth Apps → New OAuth App. Create two:
| Homepage | Callback | |
|---|---|---|
| dev | http://localhost:8788 |
http://localhost:8788/callback |
| prod | https://mcp.alanko.dev |
https://mcp.alanko.dev/callback |
No scopes needed (only the public profile is read).
2. Lunch Money
Create a personal access token at my.lunchmoney.app/developers.
3. Local dev
pnpm install
cp .dev.vars.example .dev.vars # fill in dev OAuth app + token; COOKIE_ENCRYPTION_KEY = openssl rand -hex 32
pnpm dev # http://localhost:8788/mcp
pnpm test
4. Deploy
npx wrangler kv namespace create OAUTH_KV # put the returned id into wrangler.jsonc
pnpm deploy
# production secrets:
npx wrangler secret put GITHUB_CLIENT_ID
npx wrangler secret put GITHUB_CLIENT_SECRET
npx wrangler secret put COOKIE_ENCRYPTION_KEY
npx wrangler secret put ALLOWED_GITHUB_LOGINS # your GitHub username
npx wrangler secret put LUNCHMONEY_TOKEN
CI (.github/workflows/ci.yml) typechecks/lints/tests every PR and push to main, and deploys on main after checks pass. Repo secrets required: CLOUDFLARE_API_TOKEN (Edit Workers template), CLOUDFLARE_ACCOUNT_ID.
5. Connect to Claude
claude.ai → Settings → Connectors → Add custom connector → https://mcp.alanko.dev/mcp → sign in with GitHub.
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。