[FEATURE] Per-text color customization within AI responses

Preflight Checklist

• [x] I have searched existing requests and this feature hasn't been requested yet
• [x] This is a single feature request (not multiple features)

Problem Statement

Summary

Add support for customizing the color of specific text patterns within Claude Code's AI response output — for example, allowing users to assign a distinct color to inline code, foreign language phrases, or other structured text elements.

Use Case

I use Claude Code with a French language learning practice embedded in my workflow: Claude includes French phrases and pronunciation guides inline in responses. Currently, all response text renders in the same color, making it difficult to visually scan for the French elements without reading every line.

Being able to configure a distinct color (e.g., via a settings.json rule or a theme extension) for matched text patterns would make this kind of embedded practice significantly more effective.

What I've already explored

• Changing my Windows Terminal color theme — this affects the whole terminal, not response text
• Claude Code's /config theme options — applies to the UI chrome, not response text content
• ANSI escape codes — Claude Code does not expose these in prose output

Related Issues

• #34172 — Customizable colors/background for user input vs AI output (closest match, but covers global input/output differentiation, not per-text patterns within a response)
• #34121 — Persist prompt color across sessions (prompt-level, not response-text-level)

Environment

• OS: Windows 11
• Terminal: Windows Terminal
• Claude Code version: latest

Proposed Solution

Proposed Solution (one or more of)

1. Named style tags: Allow Claude to emit lightweight markup (e.g., {fr}bonjour{/fr}) that the renderer maps to a user-configured color

1. Semantic text class coloring: Expose existing markdown element types (inline code, blockquote, etc.) as color-configurable in settings

1. (Claude Code suggested this; but as I read it, RegEx matching would be pretty frail - but ANSI colours could be a solution)

Pattern-based coloring in settings.json: Allow regex rules that apply an ANSI color to matched text in assistant responses
``json
"responseTextColors": [
{ "pattern": "\\[.*?\\]", "color": "cyan" }
]
``

Alternative Solutions

_No response_

Priority

Low - Nice to have

Feature Category

Configuration and settings

Use Case Example

_No response_

Additional Context

_No response_