Scale Intelligence■ Docs
Docs / Reference / For models

For models.

A language model reads this site through the index for models, the full Markdown copy, the Markdown twin of every page, the API document, the MCP manifest and the Context7 configuration.

Figure 1Reading path for modelsspecified
MODELreads text insteadof rendered pages/LLMS.TXTevery page as a linkone line eachMARKDOWN TWINthe address with .mdsame front matterThe full Markdown copy at /llms-full.txt holds every page in reading order.flows forwardthe result
A model reads the index for models first, then the Markdown twin of each page that it needs.
file address contents
the index for models /llms.txt a heading with the platform’s name, a paragraph about the platform, then the sections, and every page as a link with a one-line description
the full Markdown copy /llms-full.txt every page’s Markdown, in reading order
the Markdown twin the page’s address with .md appended the page’s Markdown, with the same front matter, headings and examples
the API document /openapi.json the served OpenAPI document at contract 0476e35e6e275db5, byte for byte
the MCP manifest /mcp-manifest.json the served manifest at contract 0476e35e6e275db5, byte for byte
the Context7 configuration context7.json at the repository’s root the docs folder and its rules for readers
the MCP resource docs://{address} on the MCP server the page’s twin, which the MCP server serves to a connected agent without a key