Open this page in the study app
Use this site with an AI assistant
This site has an MCP server (Model Context Protocol): an AI assistant connected to it studies from the same published content as you, with the source of every item, and can run quizzes, mock exams, flashcards and trade-math drills whose answers are revealed only after you answer. No account, no key, nothing stored.
Server address
The server runs on this site at the address below (Streamable HTTP transport, no authentication):
https://cabinetmaker-red-seal-prep.pages.dev/mcp
What the assistant gets
- The study content in English and in French (two distinct content sets, each written from official sources in its own language), the occupational standard, the exam information, the glossary, the reference library and the source registry, with the citation of every item.
- Practice: quizzes with the official exam weighting, the full mock exam with the official allocation and pass mark, flashcards and generated trade-math drills, in the wording of this site.
- The exam-province filter of this site: Canada-wide content plus the chosen province’s (New Brunswick by default).
- Ready-made study flows (a study session on a task, a mock exam, a review of mistakes) and the content by address, for clients that attach resources.
Connect an assistant
Claude.ai and Claude Desktop (custom connector)
- Open Settings, then Connectors, and choose “Add custom connector”.
- Give it a name such as “Cabinetmaker Red Seal Prep”, paste the server address, and leave the authentication fields empty.
- In a chat, turn the connector on and ask, for example: “Quiz me on task B-6, five questions” or “Summarise the study guide for task F-14 and cite the sources.”
Claude Code
In a terminal, run:
claude mcp add --transport http cabinetmaker https://cabinetmaker-red-seal-prep.pages.dev/mcp
Other MCP clients
Any MCP client that supports the Streamable HTTP transport can use the address above; there is no authentication. The server lists its tools, prompts and resources when the client connects, and the setup guide describes each of them.
Offline: run the server on your computer
The same server also exists as a single Node.js script (Node.js 20 or newer) packaged with a copy of the study content, so an assistant on your computer can use it without an internet connection. It is not published on npm.
Download the cabinetmaker-mcp-stdio archive from the latest build on GitHub (open the most recent run and look under “Artifacts”; a GitHub account is needed to download it), or build it from the source code as the setup guide explains. Unzip it and keep the content folder next to stdio.mjs. Then register it with your client, replacing /path/to/stdio.mjs with the full path of the file:
Claude Desktop: add this to claude_desktop_config.json (Settings, Developer, Edit Config), then restart Claude Desktop:
{"mcpServers":{"cabinetmaker":{"command":"node","args":["/path/to/stdio.mjs"]}}}
Claude Code:
claude mcp add cabinetmaker -- node /path/to/stdio.mjs
The offline copy contains the content published when it was built; download a new build after the site is updated. Only MCP messages go to standard output; notes and errors go to standard error.
Rules the assistant follows
- Practice questions, mock exams and drills are returned without their answers. The assistant shows each question as returned, waits for your answer, and only then asks the server to check it; the answer, the explanation and the sources come back with the feedback, as on this site.
- Every item carries its citations (source title, locator, address). The server instructs the assistant to answer only from the returned content and to cite them; repeat that in your own prompt, and check the source when it matters.
- The server is public and read-only, keeps no state and stores nothing about you: a quiz or an exam is rebuilt from its seed.
- The site and its server are an independent study aid, not affiliated with or endorsed by the Red Seal Program or any apprenticeship authority. They contain no real exam questions.
For crawlers and other tools
llms.txt describes the site and its files for AI tools. These readable pages and the sitemap publish the study content as plain HTML, and the content index lists every published file. Full documentation of the tools, prompts and resources: the MCP guide.