#!/bin/bash # ~/.mycfg/scripts/ow-dictate.sh — OpenAI Whisper API version PIDFILE="/tmp/dictate.pid" AUDIO="/tmp/dictate.wav" NOTIFFILE="/tmp/dictate.notif_id" NID="" [ -f "$NOTIFFILE" ] && NID=$(cat "$NOTIFFILE") notify() { local args=() [ -n "$NID" ] && args+=(-r "$NID") NID=$(notify-send -p "${args[@]}" "$@") echo "$NID" > "$NOTIFFILE" } if [ -f "$PIDFILE" ] && kill -0 "$(cat "$PIDFILE")" 2>/dev/null; then # ── STOP recording, send to API, type ── kill "$(cat "$PIDFILE")" rm "$PIDFILE" notify -t 2000 "Dictation" "Sending to OpenAI..." API_KEY=$(pass show openai.com/api.openai.com/apikey) RESPONSE=$(curl -s https://api.openai.com/v1/audio/transcriptions \ -H "Authorization: Bearer $API_KEY" \ -F "file=@$AUDIO" \ -F "model=whisper-1" \ -F "language=en" \ -F "response_format=json") TEXT=$(echo "$RESPONSE" | jq -r '.text // empty') if [ -n "$TEXT" ]; then xdotool type --clearmodifiers -- "$TEXT" notify -t 1000 "Dictation" "Done" else ERROR=$(echo "$RESPONSE" | jq -r '.error.message // "Unknown error"') notify -t 3000 "Dictation" "Failed: $ERROR" fi rm -f "$AUDIO" else # ── START recording ── notify -t 0 -u low "Dictation" "Recording..." arecord -f S16_LE -r 16000 -c 1 -t wav "$AUDIO" & echo $! > "$PIDFILE" fi