Thinking Claude
Prompt instructions and a Chrome extension that make Claude show more structured reasoning. Retired.
- Stars
- 17k
- Forks
- 2k
- Licence
- MIT
- Language
- TypeScript
- Last update
- 7 months ago
- Type
- Other
What it is
Thinking Claude is a set of written instructions (a "thinking protocol") that asks Claude to reason more thoroughly before answering, plus a Chrome browser extension that organizes Claude's thinking into collapsible, copyable sections on claude.ai. The README says the repository has been retired because Claude is already a reasoning model with strong built-in thinking, so the project is no longer necessary.
- Paste a chosen instruction file from model_instructions/ into a Claude custom style on claude.ai
- Install the Chrome extension to fold, unfold, and copy parts of Claude's thoughts
- Compare the versioned instruction files to see how the prompt changed over time
Great for
- People who want to read Claude's step-by-step reasoning in a neater format on claude.ai
- Researchers curious about how prompt instructions change a model's visible thinking
- Anyone looking at the historical approach of prompt-based reasoning for Claude 3.5 Sonnet
Not ideal for
- Anyone expecting current support: the README says the repository has been retired
- Users who need better scores on math or benchmarks: the README states this project is not aimed at that
- Users of Claude models other than the one referenced, since the README only mentions Claude 3.5 Sonnet
How to get it
- 1
Install the browser extension
You'll need: Google Chrome or a Chromium-based browser · A Claude.ai account (Free or Pro)
Clone the repository
git clone https://github.com/richards199999/Thinking-Claude - 2
Run it
- Download the latest Chrome extension zip from the Releases page and unzip it
- Open chrome://extensions/, turn on Developer mode, and click Load unpacked on the unzipped dist folder
- Visit claude.ai, open the style selector, and paste an instruction file from model_instructions/ as a custom style
- Start a chat and check that Claude's thinking appears in collapsible sections
- 3
Get the most out of it
- Try the extended and lite instruction versions to see how length changes the output
- Use the copy button in the extension to save useful parts of Claude's reasoning
- Read the changelog files in model_instructions/ and extensions/ to understand each version
Downloads
Strengths
- Simple to try: the core is plain text instruction files you can read
- The browser extension makes long reasoning easier to scan and copy
- MIT licensed, so the files and code can be reused and modified
Trade-offs
- Retired by its maintainers, so it is unlikely to receive updates or fixes
- The README says it targets Claude 3.5 Sonnet and makes no claims about newer models
- Its benefits are described qualitatively; the README gives no measured results
Stats synced from GitHub. The assessment above was drafted with AI from the project's README on 2026-10-29 and may miss recent changes — check the README before relying on a command. MIT licence: free to use, modify and distribute, including commercially, as long as you keep the copyright notice.