Run It on AI

Run It on AI

Read-only MCP connector that serves the Run It on AI book. The index and Implementation Blocks are free; full chapters and playbooks unlock with a license key included with the book.

Category
访问服务器

README

Run It on AI — the companion server, in full

This repository is the complete source code of the server that powers the Run It on AI reading companion at mcp.runitonai.com. If you bought the book and connected your AI to it, this is everything that runs on our side.

We publish it so you do not have to take our word for anything. The server is a few hundred lines of TypeScript. You (or your AI — paste any file into it and ask) can read exactly what it does and, just as importantly, what it cannot do.

The short version

What reaches our server when you use the companion:

  • The name of the tool your AI called (for example "fetch chapter 3").
  • The chapter or exercise it asked for.
  • Your access code — the license key printed on your Gumroad receipt.

What never reaches us, because there is no code here that could receive, store, or forward it:

  • Your conversations with your AI.
  • Your files, documents, or anything on your machine.
  • Your name.
  • Your email. Search this repository for an email field: there is none. No tool, no endpoint, no code path accepts one.
  • The picture your AI forms of you while you work. That stays inside your AI, on your side.

Who talks to whom: when you add the companion as a Claude connector, your machine never contacts our server directly — Anthropic's cloud brokers the connection. On ChatGPT, OpenAI's cloud does. We see tool calls, not you.

How access works

Your credential is the license key on your Gumroad receipt (a code like XXXXXXXX-XXXXXXXX-XXXXXXXX-XXXXXXXX). It identifies a purchase, not a person. The server checks it against Gumroad, caches a hash of it for a day, and rate-limits it. Refunded or disabled keys stop working. See src/license.ts — the whole mechanism is one file.

Only chapter text needs the key. The companion's operating instructions (activate) and every Implementation Block — the book's working exercises — are served without any credential at all; the exercises are also published right here in implementation-blocks/.

Connect it

Claude (web or desktop) — the easy path. Settings -> Connectors -> Add custom connector, then paste:

https://mcp.runitonai.com/mcp

Start a chat with: Activate Run It on AI — my access code is <your license key>.

ChatGPT. Use the "Run It on AI Companion" GPT (link in the book), or build your own: create a Custom GPT and import https://mcp.runitonai.com/openapi.yaml as an Action.

Claude Code (terminal users).

claude mcp add --transport http runitonai-book https://mcp.runitonai.com/mcp

Codex. In ~/.codex/config.toml:

[mcp_servers.runitonai_book]
url = "https://mcp.runitonai.com/mcp"

What is in this repo

Path What it is
src/worker.ts The Cloudflare Worker: routing, transports, the Gumroad sync endpoint
src/license.ts License-key verification and caching (the access gate)
src/auth.ts Verification of legacy bearer tokens (nothing mints new ones)
src/rate_limit.ts Per-key rate limiting
src/mcp/ MCP protocol plumbing (tool catalog, dispatch, transports)
src/tools/ The tools your AI can call
openapi-v1.yaml The same surface described for ChatGPT Actions
implementation-blocks/ Every Implementation Block from the book, as readable markdown

Two data files ship as placeholders: src/data/chapters.json (the book's chapter text) and src/data/skill.json (the companion's operating instructions). Both are generated from the book manuscript at deploy time. The book's prose is the product, so it is not in the public repo — but the code that serves it is all here, and the Implementation Blocks (the working exercises) are published in full in implementation-blocks/.

Note for contributors: everything in this repo is generated from the private manuscript repo by a sync script, and a pre-commit hook there re-syncs it on every relevant change. The Implementation Blocks live in the book's manuscript; the copies here are build artifacts — never edit them here, they will be overwritten.

Run it yourself

npm install
npm run typecheck
npm run dev        # local server on localhost:8787

wrangler.toml ships with placeholder KV namespace ids; point them at your own Cloudflare account if you want to self-host the shell.

License

MIT — see LICENSE. The book itself (chapters, companion instructions) remains copyrighted content of StructLabs.io Pte Ltd and is not part of this repository.

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