Source Intelligence

DisclaimerUnofficial, and not affiliated with Anthropic. Nearly all of this is read straight out of what ships: npm bundles, captured prompts, published docs. Anthropic's own notes go in verbatim, marked as theirs. The rest is my reading, and every entry carries the strings behind it. If one looks wrong, vote it down and say why.

All of v1.0.114 Home All releases olderv1.0.113 v1.0.115newer
Claude Code v1.0.114

Interactive Thinking Mode with Visual Highlighting

What: Real-time detection and visual highlighting of thinking trigger words in 7 languages (English, Japanese, Chinese, Spanish, French, German, Korean) with three intensity levels.

How to use: Type trigger words in your prompt to activate extended thinking:

# Basic (4,000 tokens): "think", "考えて", "想"
claude "think about the best algorithm for this"

# Medium (10,000 tokens): "think deeply", "megathink", "もっと考えて"
claude "think deeply about the edge cases"

# High (31,999 tokens): "think harder", "ultrathink", "熟考"
claude "think harder about the security implications"

Details:

  • Trigger words are highlighted with shimmer effects as you type
  • Status bar shows: "Thinking on • [level] • /t to disable"
  • Use /t command to toggle thinking mode on/off
  • Intelligently ignores conversational phrases like "I think" and "we think"
  • Color-coded by intensity: basic (low), medium (medium), high (max)
  • Evidence: TTA() at line 366061, PTA() at line 366112, fA0() at line 366083, o1Q() at line 436378

Related

Other releases about the same thing. Found by shared names or similar wording; neither means one caused the other.

See this entry in the whole of v1.0.114 →