Skip to main content
These docs are built to be read by AI assistants and agents as much as by people. If you’re pointing a model at Mirrors (or you are the model), here’s the fast path.

​
Raw markdown for every page

Append .md to any docs URL to get the page as clean markdown, no HTML or chrome:

​
The index: llms.txt

One file lists every page with its description. Fetch it first to decide what to read:
There’s also a product-level overview at /llms.txt (positioning, pricing, key facts), and llms-full.txt bundles the entire docs into a single file if your tool prefers one fetch.

​
One-click handoff

Every page has a contextual menu in the top-right corner:
  • Copy page: the page as markdown, straight to your clipboard.
  • View as Markdown: the raw .md twin.
  • Open in ChatGPT / Claude: starts a chat with the page already loaded as context.
  • Open in Cursor / VS Code: pulls the page into your editor’s AI chat.
You can also ask questions right here: the docs’ built-in assistant answers from this documentation and links its sources.

​
Drive the product itself over MCP

Mirrors is also operable by agents, not only readable. The hosted MCP server exposes the full product surface (build, explore, query, and eval environments), so an assistant that has read these docs can do the work too:
A good first prompt for an agent: “Fetch https://www.runmirrors.com/docs/llms.txt, read the pages relevant to my task, then connect to the Mirrors MCP server and do it.”