Taskriz 远程 MCP 服务器是什么?
Taskriz 的远程 MCP 服务器是一个托管的 Model Context Protocol 端点,地址是 https://www.taskriz.com/mcp,AI 助手无需在本地安装或运行任何东西即可连接。

@Appeggio.com Trading Pty Ltd
基本信息
Taskriz lets your AI agent ask a person for a decision, a date, a photo or a document, and get a structured answer back. It sends the person a screen made for the request, by email, text message or the Taskriz app, and returns their answer to your agent. Works with Claude, ChatGPT and any MCP client. Free to start.
https://www.taskriz.com/mcpclaude mcp add taskriz --transport http https://www.taskriz.com/mcp工具列表暂未提供
当前收录暂时没有可展示的工具列表,不代表服务没有工具。请查阅项目文档了解支持的能力。
此服务需要认证,获取工具列表可能需要登录或使用服务提供方的凭据。
点击上方按钮,直接连接服务器端点实时获取工具列表。
Let your AI agent ask a person for a decision, a date, a photo or a document, and get a structured answer back.
Taskriz is a remote MCP server for human-in-the-loop work. Your agent says what it needs from someone, in plain words or as an interface it designs. Taskriz sends that person a screen made for the request (buttons, a date picker, a camera, a file picker, a map of the address) by email, text message or the Taskriz app, and returns their answer to your agent as structured data.
The people you ask need no app and no account: they answer from the link in the email or text message.
| Tool | What it does |
|---|---|
list_agents / create_agent | Choose or create the Agent (the sender name people see) |
list_recipients | The people this Agent can reach |
get_ui_guide / check_ui | Design a request's interface, and check it before sending |
send_request | Send a person a request, by email, text message or the app |
get_replies / acknowledge | Read the answers, and mark them as handled |
recall_request | Withdraw a request that hasn't been answered |
create_attachment / get_attachment | Send files with a request, and read files people send back |
set_branding | Your business's name, logo and footer on every request |
set_catalog | Actions people can start with your Agent ("Book a service") |
pairing_code | Connect a person's Taskriz app to your Agent |
Endpoint: https://www.taskriz.com/mcp (Streamable HTTP, OAuth sign-in)
/plugin marketplace add appeggio-com/taskriz-claude-plugin, then /plugin install taskriz@taskriz{
"mcpServers": {
"taskriz": {
"type": "http",
"url": "https://www.taskriz.com/mcp"
}
}
}
Requests are sealed so only the devices of the person you ask can open them, notices never include what was asked, and everyone chooses who can reach them.
Free to start: 500 requests a month, no credit card. Paid plans from US$29 a month. The people you ask never pay.
Taskriz 的远程 MCP 服务器是一个托管的 Model Context Protocol 端点,地址是 https://www.taskriz.com/mcp,AI 助手无需在本地安装或运行任何东西即可连接。
把端点 https://www.taskriz.com/mcp 添加到任意支持 MCP 的客户端,比如 Claude Code、Cursor 或 VS Code。本页的接入代码块可以一步配置好每个客户端。
需要。Taskriz 使用 OAuth——第一次连接时,你的 MCP 客户端会打开浏览器窗口登录并授权,之后会复用这个凭证。
Taskriz 暴露的是 Streamable HTTP 端点,这是远程 MCP 服务器普遍使用、主流客户端都支持的传输方式。
Remote MCP server that turns Claude, ChatGPT, or Cursor into a full ad workflow: research any brand's live Facebook and TikTok ads, clone the proven winners for your own brand, and launch real campaigns on Google, TikTok
CVE prioritization using exploitation signals, plus agent heartbeat monitoring and state capsules. Some tools use API-key or x402 access.
Read-only, source-backed AI model pricing, history, comparisons and cost calculations.
Search and read current AWS documentation, check regional availability and load AWS agent skills, with no AWS account needed.
Search and read public data from X/Twitter, Reddit, YouTube, Instagram, TikTok, Facebook, Threads, Bluesky, Trustpilot, Amazon, Google Maps, news, forums and the web. 90+ read-only tools plus ready-made research playbook
评论