uploadthing-mcp

uploadthing-mcp

MCP server for UploadThing that lets AI assistants upload, list, and delete files on UploadThing's CDN via natural language. Runs as a Cloudflare Worker for always-on serverless access.

Category
访问服务器

README

uploadthing-mcp

MCP server for UploadThing, deployable as a Cloudflare Worker.

Connect any MCP-compatible AI assistant (Littlebird, Claude, Cursor, etc.) to your UploadThing account. Upload files from remote URLs, list your files, and delete them — all via natural language.

Deploy to Cloudflare Workers


Why this exists

Most AI tools generate images and files at temporary URLs that expire. This MCP server lets your AI assistant permanently host those assets on UploadThing's CDN — so you can embed them in READMEs, docs, or anywhere else without worrying about broken links.

It runs as a Cloudflare Worker (serverless, always-on, free tier covers most personal use), and connects to any MCP client via a plain HTTPS URL — no local process required.


⚠️ UploadThing API version notice

UploadThing ships breaking changes to their SDK fairly frequently. This project currently tracks uploadthing@^7.x.

If you hit errors after a fresh install or upgrade:

  1. Check the UploadThing changelog
  2. Check the open issues in this repo
  3. Pin a specific working version in package.json if needed

PRs to update compatibility are very welcome.


Tools

Tool Description
upload_from_url Download a file from any public URL and re-upload to UploadThing. Returns a permanent CDN URL at https://<app>.ufs.sh/f/<key>.
list_files List files in your UploadThing app with keys, URLs, names, and sizes. Supports limit and offset for pagination.
delete_files Delete one or more files by their file keys.

Deploy

One-click (Cloudflare)

Click the Deploy to Cloudflare Workers button above. After deployment:

  1. Go to Cloudflare Dashboard → Workers & Pages → your new uploadthing-mcp worker
  2. Navigate to Settings → Variables & Secrets
  3. Add a secret: UPLOADTHING_TOKEN = your UploadThing API token
  4. Optionally add: AUTH_TOKEN = a bearer token of your choosing to protect the endpoint

Manual

Prerequisites: Node.js 22+, Wrangler CLI, an UploadThing account

git clone https://github.com/ppatel26/uploadthing-mcp
cd uploadthing-mcp
npm install

# Set your UploadThing token (Dashboard -> API Keys)
npx wrangler secret put UPLOADTHING_TOKEN

# Optional: protect your endpoint
npx wrangler secret put AUTH_TOKEN

npm run deploy

Your server will be live at:

https://uploadthing-mcp.<your-cf-subdomain>.workers.dev

Connect to your MCP client

Littlebird

Go to Settings → Integrations → Add MCP Server and paste your Worker URL.

Claude Desktop / Cursor

These clients use stdio transport and run MCP servers as local processes. For local use, see the original Toolbase-AI/uploadthing-mcp which inspired this project.

Other HTTP-based MCP clients

Point the client at your Worker URL. If AUTH_TOKEN is set, configure the client to send Authorization: Bearer <your-token> with each request.


Environment variables

Variable Required Description
UPLOADTHING_TOKEN ✅ Yes Your UploadThing API token (sk_live_...). Get it from the dashboard.
AUTH_TOKEN Optional A bearer token to protect the endpoint. If set, all requests must include Authorization: Bearer <token>.

Local development

# Create a .dev.vars file (gitignored) with your secrets
echo 'UPLOADTHING_TOKEN=sk_live_...' > .dev.vars

npm run dev
# Worker runs at http://localhost:8787

Credits

Inspired by Toolbase-AI/uploadthing-mcp, which implements the same idea as a local stdio server.


License

MIT

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