Use these docs with your AI

Every page on docs.markdrop.in is also plain Markdown, so you can hand it to an AI assistant without copying text out of a web page. Use it to ask questions about a page, or to give a coding agent the exact steps it needs.

Copy or open a page

Each page has a Copy page button next to its title. Its menu has:

Item	What it does
Copy page as Markdown	Copies the page's Markdown to your clipboard.
Copy page with prerequisites	Copies the page together with the pages it lists as prerequisites, so an assistant has everything it needs in one paste.
View as Markdown	Opens the page's Markdown in the browser.
Open in Claude	Opens claude.ai with a prompt to read this page.
Open in ChatGPT	Opens ChatGPT with the same prompt.
Open in Perplexity	Opens Perplexity with the same prompt.
Add docs to Cursor	Installs the docs MCP server in Cursor.
Add docs to VS Code	Installs the docs MCP server in VS Code, for Copilot Chat.

The prompt is "Read {page}.md and answer my questions about it." When you open an assistant, the page's Markdown is also copied to your clipboard, in case the assistant can't open the link: paste it into the chat.

To copy one section, hover over its heading and choose Copy section.

Get any page as Markdown

Add .md to a page's address:

text

https://docs.markdrop.in/cli/upload-and-replace.md

Tools that ask for Markdown get it at the normal address too. Send Accept: text/markdown:

bash

curl -H "Accept: text/markdown" https://docs.markdrop.in/cli/upload-and-replace

Give an agent all the docs

File	What's in it
`/llms.txt`	An index of every page, with a one-line description and a link to its Markdown. Small enough to keep in an agent's context and follow links from.
`/llms-full.txt`	Every page in one Markdown file. Use it when the agent should have everything at once.
`/changelog.md`	The changelog as Markdown, newest first.

For example, in Claude Code:

text

Read https://docs.markdrop.in/llms.txt, then use the markdrop CLI to publish build/demo.mp4 into my Acme workspace's releases folder.

Ask the docs

Ask AI answers questions from these docs, with a numbered link to the exact section each part of the answer came from. Choose Ask AI in the header, press ⌘I, use Ask about this next to any heading, or press ⌘K and then Tab. Selecting a source scrolls to that section and highlights it.

It only answers from these docs. When they don't cover something, it says so instead of guessing.
It can't change anything in your account; it explains how to do things.
What you type is sent to Ask AI only when you ask. Searching with ⌘K never leaves your browser.
There is a limit of 20 questions an hour per person. If Ask AI is busy or off, you get links to ask Claude or ChatGPT with the page attached instead.

Connect the docs to your AI app

These docs have their own MCP server at https://docs.markdrop.in/mcp. Connect it, and your AI app can search and read the docs as tools whenever it needs them, without you pasting anything. It is public and read-only: no account and no sign-in.

Add to Cursor opens Cursor and asks you to install it. Or add this to ~/.cursor/mcp.json, or to .cursor/mcp.json in one project:

mcp.json

{
  "mcpServers": {
    "markdrop-docs": {
      "url": "https://docs.markdrop.in/mcp"
    }
  }
}

Add to VS Code opens VS Code and asks you to install it. Or, for Copilot Chat in agent mode, add this to .vscode/mcp.json in your workspace:

.vscode/mcp.json

{
  "servers": {
    "markdrop-docs": {
      "type": "http",
      "url": "https://docs.markdrop.in/mcp"
    }
  }
}

In your terminal:

bash

claude mcp add --transport http markdrop-docs https://docs.markdrop.in/mcp

Add --scope user to make it available in every project.

On claude.ai and in Claude Desktop:

Open Settings → Connectors.
Choose Add custom connector, name it Markdrop docs and paste https://docs.markdrop.in/mcp.
Save it. There is nothing to sign in to.

Any app that connects to remote MCP servers over Streamable HTTP works. Add https://docs.markdrop.in/mcp with no authentication.

Add docs to Cursor and Add docs to VS Code are also in every page's Copy page menu.

Tools

Tool	What it does
`search_docs`	Takes `query` and an optional `limit` (1 to 20, 8 by default). Returns the best-matching sections and pages, each with its title, the page it is on, a link to the exact section and a short snippet.
`read_page`	Takes a page's `path`, like `/workspaces/custom-domain`, or its full address. Returns the page as Markdown. With `include_prerequisites`, the pages it builds on come first.
`whats_changed`	Takes `since_date` (`YYYY-MM-DD`) or `since_cli_version` (like `0.4.0`). Returns the changelog entries since then, newest first, with the pages each one changed. With neither, returns the latest 10.
`report_problem`	Takes a `page` and the `problem`. Tells the docs team a page is wrong, out of date or missing something.

The API reference is included: read_page on /api/documents/publish-a-document returns that endpoint's parameters, request body, responses and code samples.

For example, in Cursor:

text

Using the markdrop-docs tools, find out how to replace a published PDF without changing its link, then do it for build/report.pdf.

This is separate from Markdrop's own MCP server, which works with your documents and files. To connect that, see Connect an AI app.