Linking and consuming our docs
We welcome bots, search engines, SEO tools, LLMs and other clients to use our documentation. Read a page in your browser, fetch its Markdown, discover pages through a manifest, or link to documented objects from another site.
Choose an export¶
The page footer links to the exports for its documentation tree. Use those links to select a language, version or preview.
| File or format | Use it for |
|---|---|
| llms.txt¶ | Discover guides, API references and agent prompts through a Markdown index. |
| llms-full.txt¶ | Fetch guides and MCP contracts together as Markdown. API references are linked separately. |
| docs.json¶ | Discover page titles, URLs, descriptions, headings and Markdown URLs in a structured manifest. |
| Markdown¶ | Read one page as text using its footer’s Markdown link. |
| Raw source¶ | Inspect the source file using its footer’s Raw source link, where available. |
| objects-full.inv¶ | Resolve Sphinx references across shared docs, each port’s default version, or its latest until the default has published a catalog, and tmux reference versions. |
| objects.inv¶ | Resolve Sphinx references within one port and version, or one tmux documentation tree. |
| targets.json¶ | Browse inventory keys, display names and destination URLs as JSON. A scoped catalog sits beside each root inventory. The root one lists fallbacks and ports left out. |
The root exports above cover the assembled site. Port exports describe the
selected port and version. For a single API declaration, start with
docs.json and follow its reference and symbol-index links.
Link from Sphinx¶
The Intersphinx guide explains how to configure inventories, choose native roles or portable labels, and preserve references when a page moves or changes its title.
Browse Intersphinx examples for copyable RST and MyST references to library APIs, tmux C declarations, hooks, format variables, CLI commands and schema fields, paired with their destinations.
Discover and index pages¶
Use the sitemap to discover indexable pages. Follow
robots.txt and each page’s robots metadata:
pull-request previews, component demos and search-index shards are excluded
from crawling, and alternate versions can carry noindex.
Choose a versioned documentation tree when a reference needs a particular
release. The full inventory follows each port’s advertised default; a
port’s objects.inv describes the selected version.