cc-stretch

Claude writes a lot. Now you can read a little.

Replies arrive short. Click a linked phrase and it grows right where it stands. Click « and it shrinks back. Claude is told which parts you opened, so it doesn't assume you read the rest.

In a Claude Code session:

/plugin install cc-stretch --marketplace pro-vi/cc-stretch
/tui fullscreen

In the terminal, clicks need Claude Code's fullscreen renderer. /tui fullscreen switches to it, restarting and resuming your session. If stretches still show open, see Limits. In the Claude desktop app stretches are links: click one to open it in place, click « to close it. No setting needed.

To update, run claude plugin update cc-stretch@cc-stretch in a shell, or open /plugin, find cc-stretch under Installed and choose Update now. As of Claude Code 2.1.302 a plugin from this marketplace updates on its own only once you turn on auto-update for the marketplace in /plugin's Marketplaces tab.

You read about a sixth of what Claude writes. Across 69 real replies from four of my coding projects, the version you see before clicking was 17% of the words Claude wrote without the plugin. It carries about two of every three most important points; the rest is one click away. My replies run long, so if yours are short you will save less. How this was measured, and how to run it on your own transcripts: bench/.

In 1967 Ted Nelson described stretchtext: text that "stretches, becoming longer, with replacement phrases, new details and additional clauses popping into place", and shrinks back on demand, while "the reader remains oriented" (Hypertext Note 8). In 1970 he put it like this: "Gaps appear between phrases; new words and phrases pop into the gaps, an item at a time."

It never caught on. The hard part, as later writers put it, was "how to author the text in the first place" (Wadge and schraefel, 1999): someone had to write every level by hand.

Language models write every level for free, and they are good at it. So all this plugin does is plug a 59-year-old idea into a terminal. The idea is Nelson's. Most of the code is Claude's. I mostly clicked things.

Here is what that looks like. The reply opens with

A monad is a type that wraps values and lets you chain steps that each return a wrapped value.

and one click on that sentence turns it into this (new words in bold, the way Dream Machines appears to set the words Nelson's own example adds):

In programming, a monad is a generic type that wraps values, and it lets you chain steps that each take a plain value and return a wrapped value**, without unwrapping anything by hand between the steps.**

The words you already read never move. New ones open up around them, on both sides.

• skim: a reply with every stretch closed. What you read by default.
• full: the same reply with every stretch open. Everything Claude wrote.
• read: what you actually read, the skim plus the stretches you opened.

The point of the plugin is the gap between them. Nelson's word for this was levels; ours is progressive attention. Progressive disclosure is the interface deciding what to hide from you. Here you spend your attention one phrase at a time, and because nothing you have read moves, you never lose your place.

• Stretches nest. A long form can hold more stretches.
• Words to click before a lone period. When Claude makes a sentence's period the link to more, or a spot with no words at all, the sentence's words before it open it too, marked like the link, so there is a phrase to click rather than one cell. This is in the terminal's fullscreen layout; in the Claude desktop app only the short form is the link.
• Streaming stays clean. While a reply streams, stretches already show as their short forms.
• Letters for the keyboard. In the terminal's fullscreen layout, alt+↓ or ctrl+↓ puts a letter over each closed stretch in the latest three replies on screen that hold one (up to 26), and typing that letter opens it. alt+↑ or ctrl+↑ puts letters over each «, to close. Typing anything else puts the letters away and is typed as usual. Recalling a prompt (from Claude Code 2.1.300), sending one, or pressing the same chord again puts them away too. As of Claude Code 2.1.301 a plugin has no chords of its own, and these are the ones that scroll the diff panel's file list, so while that panel is open with more than eight files they scroll it instead. As of macOS 26, ctrl+↑ opens Mission Control and ctrl+↓ shows the app's windows by default. If your terminal or system takes both chords of one action, bind another to app:diffFileListDown or app:diffFileListUp in the Global context of ~/.claude/keybindings.json.
• The marking is yours to pick. Phrases you can open are underlined by default. In the terminal's fullscreen layout, run /cc-stretch:style and a toolbar above the prompt restyles every reply as you click: marks (bold, italic, underline, strikethrough, dim) and a highlight, either one hex colour or tinted, which highlights each closed stretch in the colour its added words are drawn in once you open it, blended 75% into the background. Under a highlight, dim moves the text toward the highlight. save keeps the style in two plugin settings, marks and highlight, which as of Claude Code 2.1.302 /plugin's Configure options also change; close puts the toolbar away and drops a style you tried but did not save. In the Claude desktop app, where cc-stretch draws stretches as links, there is no toolbar.

• Clicks work in the terminal's fullscreen layout and in the Claude desktop app. Everywhere else every stretch shows open, so nothing is hidden: in other apps, and on the terminal's main screen, where a printed reply can't change. There the first such reply says why. As of Claude Code 2.1.300, some setups start on the main screen even with fullscreen set: tmux -CC, SSH from Windows, or a machine where an earlier fullscreen start failed. Starting Claude Code with CLAUDE_CODE_NO_FLICKER=1 forces fullscreen there.
• In the terminal the plugin draws each paragraph that holds a stretch itself, and as of Claude Code 2.1.302 its transcript selects nothing in it. A drag there is the plugin's: the text it crosses is drawn inverse and copied when the button comes up, and it stays within that paragraph. A phrase you can open acts when the button comes up too, so a drag that starts on one selects. A double-click there selects and copies the word under it, as Claude Code 2.1.303 does in a paragraph without a stretch, except that the # of an id such as #12806 stays in the word and any full stops after it stay out. In the words before a lone period, a double-click copies a word as well, so a single click there opens its stretch half a second later, once no second click can follow. A ctrl- or alt-click on an http: or https: link in such a paragraph opens it, as Claude Code 2.1.304 does in a paragraph without a stretch; a cmd-click, where the terminal keeps one, is the terminal's own.
• Claude writes every level up front, so a reply costs about as many tokens as before. You read less; you don't pay less.
• The {{short|long}} markup shows as raw text wherever the plugin isn't loaded.
• Built on Claude Code's function-hook plugin API, which is early access and changes between releases. Checked in the terminal on macOS with Claude Code 2.1.300 to 2.1.302 and 2.1.306, and with 2.1.292 to 2.1.294 in cc-stretch 0.1.0; in the Claude desktop app with 2.1.295 in cc-stretch 0.2.0. Claude Code before 2.1.300, which the desktop app can still bundle, refused cc-stretch 0.4.0 to 0.5.2 whole (#17); 0.5.3 loads in 2.1.293 and 2.1.295 again, but its clicks in the desktop app there are unchecked. On Linux, a reader reports clicks working with 2.1.302 in a terminal pane of the Orca app (#14). On Windows 11, a reader reports clicks working with 2.1.304 in PowerShell (#1).
• A terminal session that also has the desktop app attached shows the markup while a reply streams, because the desktop builds its links from it.

npm test            # claude plugin validate . && claude plugin test .
npm run typecheck   # the benchmark scripts

The launch film in film/ is made from a real recorded session: film/take.mjs drives Claude Code in a WezTerm pane and records its screen, and film/compose.mjs renders the film. Each release from 0.4.0 on has a film of its own: film/releases/<version>/ holds its take script and cut, and node film/compose.mjs --release=<version> renders it. Archive pictures in the launch film: photographs of the Hypertext Editing System at Brown University, October 1969, by Gregory Lloyd (CC BY-SA 4.0, via Wikimedia Commons), and a photograph of Ted Nelson, 2010, by Beatrice Murch (CC BY 2.0, via Wikimedia Commons). They are not stored in this repo; film/CREDITS.md lists each source, its license and where to fetch it into film/out/assets/.

Not affiliated with Anthropic. Clawd belongs to Anthropic; stretchtext belongs to Ted Nelson.