Skip to main content
An HTML document is a formatted report, dashboard, or visual summary that a teammate opens by link. An agent builds it, and it renders as a page.
HTML is a file type, not a render format. A markdown document can carry a render format (kanban, tracker, table, base) and stays markdown underneath. An HTML document is not markdown at all, and never carries a render format.

​
How one gets created

Only a connected agent can create an HTML document. There is no way to author or edit one in the web app: no new-HTML-document button and no editor for them. Ask your agent for “a one-page HTML report of Q3 channel performance” and it writes the document; ask it again to change something and it replaces the body, which is checked against the same rules as the original.

​
What the HTML must be

A complete, self-contained, static HTML document. The document renders inside a sandbox that does not execute scripts and blocks every network request. A script will not run, and a stylesheet, font, or image fetched from somewhere else will not load — the page renders without it. Ordinary hyperlinks are fine: the restriction is on resources the page needs to fetch in order to render. Tatara checks all of this at write time and rejects HTML that breaks the rules, telling the agent exactly what to fix. A document that lands in your brain is one that renders.

​
How it renders for a reader

An HTML document is read-only. It fills the reading area edge to edge, with no title heading, no width control, and no editor. Two actions sit in the topbar:
  • Open raw in new tab opens the file on its own.
  • Expand gives it the full window.
Unlike a document that carries a render format, an HTML document does have the properties rail, and it gains a File card at the top showing the filename, the line embedded · static html · no scripts · sandboxed, and an Open raw ↗ link. If the content cannot be fetched, the page shows Couldn't load HTML content. with a Retry button.