Usher MCP

Usher MCP

Enables users to search and view detailed movie information from TMDB, including cast, ratings, and showtimes, through an interactive widget interface.

Category
访问服务器

README

Usher MCP — Movie Detail App

Interactive MCP server and UI widget that renders movie details from TMDB with the MCP Apps spec. The server runs on Cloudflare Workers; the UI is a sandboxed HTML resource served via ASSETS and rendered by MCP hosts.

What’s here

  • MCP server (server/) exposing:
    • ui://widget/movie-detail-widget.html resource (HTML + CSS + JS from ASSETS)
    • Tool get-movie-detail that searches TMDB by title and returns details + cast
  • Widget UI (web/) built with React/Tailwind:
    • Entry: widget.htmlmovie-detail-widget.js/.css
    • MCP Apps lifecycle: ui/initialize, tool input/result notifications, ui/open-link for showtimes
    • Showtimes link opens Google “<title> showtimes near me” via ui/open-link
  • Root landing page: friendly message at / telling visitors to connect via /mcp

Prerequisites

  • Node 18+
  • Wrangler CLI
  • TMDB API token (v4 bearer) stored as a secret

Install

npm install

Local dev (Workers)

Use .dev.vars for local secrets:

TMDB_TOKEN=your_tmdb_bearer_token

Run dev server:

npm run dev

MCP endpoint will be at http://localhost:8787/mcp.

Build the widget

npm run web:build

Outputs to web/dist/:

  • movie-detail-widget.js
  • movie-detail-widget.css
  • widget.html

These are served by the ASSETS binding for the MCP UI resource.

Deploy to Cloudflare Workers

  1. Set the secret in your account:
wrangler secret put TMDB_TOKEN
  1. Deploy:
npm run deploy

MCP server code pointers

  • server/lib/mcp.ts — registers resource + tool, uses ASSETS to serve built widget, pulls tmdbToken from env binding.
  • server/index.ts — Hono entry, mounts /mcp and root landing page.
  • web/src/movie-detail-widget.tsx — MCP Apps UI logic + render.
  • web/vite.config.ts — builds widget assets with custom filenames for ASSETS.

MCP tool contract

  • Tool: get-movie-detail
  • Input: { "query": "<movie title>" }
  • Returns:
    • content: text fallback
    • structuredContent.movie: movie payload (title, poster/backdrop URLs, runtime, genres, rating, cast, etc.)

Notes for hosts

  • UI resource URI: ui://widget/movie-detail-widget.html
  • MIME: text/html+mcp
  • CSP: allows https://image.tmdb.org/ for posters/backdrops
  • Expects MCP Apps messages:
    • ui/notifications/tool-input / tool-input-partial
    • ui/notifications/tool-result
    • ui/open-link (to open showtimes link)

Type generation

Regenerate Cloudflare binding types if config changes:

npm run cf-typegen

Troubleshooting

  • Missing TMDB token: set TMDB_TOKEN in .dev.vars (local) and via wrangler secret put (prod).
  • Assets not loading: ensure npm run web:build was run and ASSETS binding is present.
  • Host rejects ui/open-link: host must support MCP Apps ui/open-link; otherwise widget will show an error state.

推荐服务器

Baidu Map

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。

官方
精选
JavaScript
Playwright MCP Server

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。

官方
精选
TypeScript
Magic Component Platform (MCP)

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。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

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。

官方
精选
本地
TypeScript
VeyraX

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。

官方
精选
本地
graphlit-mcp-server

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。

官方
精选
TypeScript
Kagi MCP Server

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。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

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。

官方
精选