Gist

Last updated 2026-04-19 · 1 min read

The Gist is cut off or wrong

Truncated or off-target Gists almost always mean the source was partial, dynamic, or outside the first-class content types. Try /try with the full text as the cleanest fallback.

A wrong-feeling Gist is almost always one of three things.

1. The source was partial

Paywalls, "click to continue reading" walls, and dynamically-loaded content can all leave Gist reading only the visible part of the page. The Gist is correct for what it saw. It is not what you expected it to see.

2. The source is not first-class

Regular web articles, papers, PDFs, YouTube, and pasted text are the paths Gist is tuned for. Google Docs and X/Twitter are best-effort — expect more variation. See What Gist works on.

3. The source is structurally unusual

A listicle, a long interview transcript, an FAQ, a page with ten disjoint sections — none of these has a single thesis to layer. Gist still produces something. It will not feel like a clean Gist → Story → Logic → Counter-argument → Steelman read.

What to try

  • Open the page in a normal tab and scroll. If part of the article is hidden behind a "read more" wall, Gist read the visible part.
  • Reload the page and re-Gist. Some content loads after Gist first reads.
  • Try /try with the same URL. /try reaches pages differently than the extension.
  • Paste the full text into /try when you can see the article but Gist cannot. Raw pasted text is first-class.
  • Check the layer labels. The first layer should be a one-slide hook. If it is a bullet list or paragraph of narration, Gist classified the piece as story-shaped — swipe forward to Logic to see the argument.

When it is actually wrong

If Gist confidently says something the source does not support, that is a bug. Send it in through Report a bug:

  • The source URL.
  • A short quote of what the source actually says.
  • A screenshot of the incorrect Gist layer (helpful, not required).

This is the most useful feedback we get.

What is next

Related