[FEATURE] Add voice/audio output for Claude's responses in Claude Code

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

Claude Code supports voice input (dictation/push-to-talk), so I can speak my prompts, but Claude's responses are text-only. When I want a genuinely hands-free workflow - multitasking, stepping away from the screen, or just working the way Claude.ai's voice mode already allows - I still have to look back at the terminal/panel to know what Claude said. This breaks the hands-free loop that dictation is supposed to enable.

Proposed Solution

Add an optional voice/audio output mode to Claude Code (CLI and VS Code extension) that reads Claude's responses aloud - similar to the two-way voice mode already shipped in the Claude.ai app. Ideally:

• Can be toggled independently of dictation (e.g. /voice-output, or a setting)
• Runs alongside the existing text output, not a replacement for it
• Summarizes or skips large code blocks/diffs aloud rather than reading them verbatim
• Reuses the voice/language settings already in Claude.ai (Settings > Voice)

Alternative Solutions

• Piping responses through a third-party TTS service (e.g. via a custom MCP server) - works, but needs a separate API key and isn't integrated into the product
• Using the Claude.ai app's voice mode instead of Claude Code - but that loses IDE context, tool calls, and codebase awareness that make Claude Code useful in the first place

Priority

Medium - Would be very helpful

Feature Category

Interactive mode (TUI)

Use Case Example

Example scenario:

1. I'm running a Claude Code session in VS Code, working through a feature on a healthcare data platform.
2. I ask Claude to review a set of failing tests and summarize the root cause.
3. I step away from the screen for a minute while it works.
4. With voice output on, Claude reads back a spoken summary instead of me needing to return to the editor to read it — saving time and reducing context-switching over a long dev session.

Additional Context

Claude.ai already has a full two-way voice mode (tap the waveform icon) — this request is asking for output parity in Claude Code specifically, since today it only has voice input.