Voice Mode Enhancements / Claude Code v2.1.116

The voice mode system received substantial improvements:

/voice now accepts explicit arguments: [hold|tap|off] instead of just toggling on/off
Voice enabled message now specifies the mode: "Voice mode enabled (hold). ..." or "Voice mode enabled (tap). ..."
Voice settings now have a proper description: "Voice mode settings (hold-to-talk / tap-to-toggle dictation)"
Auto-submit support: voice.autoSubmit setting enables automatic message submission when hold-to-talk is released
Silence timeout: recordings auto-finish after 15 seconds of silence (configurable)
Max-duration cap: recordings auto-finish after 120 seconds to prevent runaway recordings
Voice mode selection is now dynamic from config instead of hardcoded to 'hold'

Evidence

Voice argument hint (search for "[hold|tap|off]") — line ~547157; auto-finish (search for "Toggle silence timeout") — line ~546568

Strings lifted out of the shipped bundle, so the claim above can be checked against them.

The build it came inv2.1.116

20 Apr 2026 · 34 entries, this one is #13.

Either side of it

Open the release →

Names in this entry

slash/voice
Toggle voice dictation, or enable it in a specific mode. Requires a Claude.ai account
settinghold
Task notifications from a coordinator session now arrive automatically unless you set crossSessionInbound to refuse or hold.
settingoff
The feedback tool now says it covers Claude's own behaviour, not just product bugs.
settingautoSubmit
Under voice. When true, a voice transcript inserted into the prompt is submitted right away if the transcript has at least 3 words; in tap mode this submit happens regardless of the setting. Only the literal value true enables it.

Is this right?