wispr:on

Start the Wispr Flow voice monitor for this session. Each dictation you speak will arrive as a Monitor notification.

SKILL.md

GitHub

If ~/.claude/wispr-config.yml does not exist, suggest running /wispr:setup first, then continue anyway — monitor.sh will create a default config.

Read ~/.claude/wispr-config.yml if it exists. If wispr_app is set, pass it as an env var:

WISPR_APP=<wispr_app value> bash $CLAUDE_PLUGIN_ROOT/scripts/monitor.sh

Otherwise start without it:

Monitor(
  command="bash $CLAUDE_PLUGIN_ROOT/scripts/monitor.sh",
  label="Wispr Voice"
)

Once the monitor is running:

  • Wispr Flow transcriptions prefixed with the trigger word (default: claude) arrive as notifications. The trigger word is stripped before the text reaches you.
  • Treat each notification as a spoken message from the user. Respond normally.
  • To speak a response aloud: bash $CLAUDE_PLUGIN_ROOT/scripts/speak.sh "your reply"