Voice mode: add toggle/continuous listening mode alongside push-to-talk

Status Fixed / completed
Maintainer reply None cached
Activity 6 comments · opened Mar 4, 2026 · closed Aug 17, 2026

Problem

Voice mode currently only supports push-to-talk (hold Space to record). This creates a frustrating experience for users who rely on voice input specifically because of hand/wrist strain from typing — the very problem voice mode should solve.

Holding Space continuously while speaking is arguably more taxing on the wrist than typing, since it requires sustained pressure rather than brief keystrokes. It also makes it impossible to interleave voice and keyboard input naturally — you can't glance at output, type a quick correction, then keep talking without physically managing the spacebar.

Proposed Solution

Add a toggle mode for voice input — something like:

  • /voice --toggle or a separate /talk command
  • Press a key once to start listening, press again (or hit Enter) to stop
  • Or even better: continuous listening that automatically sends transcribed text when you pause, and lets you type normally in between

The current push-to-talk could remain as the default for users who prefer it. This would just be an additional option.

Why This Matters

Voice input is most useful for people who think and communicate verbally — and who may have physical reasons to minimize keyboard use. The current implementation forces a choice between:

  1. Sustained key pressure (worse for wrists than typing)
  2. Putting a physical weight on the spacebar (which then blocks normal keyboard use)
  3. Not using voice mode at all

A toggle mode would make voice input genuinely hands-free and practical for real workflows where you naturally switch between speaking and typing.

View original on GitHub ↗

4 Comments

github-actions[bot] · 5 months ago

Found 3 possible duplicate issues:

  1. https://github.com/anthropics/claude-code/issues/29399
  2. https://github.com/anthropics/claude-code/issues/30392
  3. https://github.com/anthropics/claude-code/issues/27908

This issue will be automatically closed as a duplicate in 3 days.

  • If your issue is a duplicate, please close it and 👍 the existing issue instead
  • To prevent auto-closure, add a comment or 👎 this comment

🤖 Generated with Claude Code

antonkor · 5 months ago

Related: #33792 proposes configurable VOICE_SILENCE_TIMEOUT_MS — even with a toggle/continuous mode, pause tolerance needs to be configurable for slower speakers. These two features complement each other.

DeanBaron · 5 months ago

Definitely much needed. Voice performs great but needing to hold space is very cumbersome. I switch between windows often. for longer prompts its really bad UX.

mbailey · 2 months ago

VoiceMode (voicemode.dev) is touch free (and screen optional).

If you just want dictation, you can disable TTS but if you want to continue working with your thinking partner in the shower or on the road, default VoiceMode might be a better option. Airpods are my primary interface to Claude Code (laptop in backpack on or passenger seat with lid closed - clamshell mode).

  • Free and open source
  • Fully configurable
  • Mature: 12 months, 1.2k 🌟, 250k PyPi downloads

https://voicemode.dev

Showing cached comments. Read the full discussion on GitHub ↗