blob: dbf89b6d0ebe110de3eeeac5afa0e43d6b1004a1 (
plain) (
blame)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
|
#!/usr/bin/env sh
# Push-to-talk dictation. Toggle: 1st invocation starts recording;
# 2nd stops, transcribes via whisper.cpp, types the result into the focused
# window with wtype, and copies it to the clipboard.
#
# Pacman: pipewire-pulse (parecord), wtype, wl-clipboard, libnotify
# AUR: whisper.cpp, whisper.cpp-model-large-v3-turbo-q5_0
#
# Override via env:
# WHISPER_MODEL path to a ggml-*.bin model
# WHISPER_LANG language code or 'auto' (default: auto)
set -eu
state_dir="${XDG_RUNTIME_DIR:-/tmp}/dictate"
pid_file="$state_dir/pid"
wav_file="$state_dir/audio.wav"
log_file="$state_dir/whisper.log"
default_model='/usr/share/whisper.cpp-model-base/ggml-base.bin'
model="${WHISPER_MODEL:-$default_model}"
lang="${WHISPER_LANG:-auto}"
mkdir -p "$state_dir"
is_recording() {
[ -r "$pid_file" ] && kill -0 "$(cat "$pid_file")" 2>/dev/null
}
start_recording() {
if ! command -v whisper-cli >/dev/null; then
notify-send -u critical "🎙️ dictate" \
"whisper-cli not found. Install whisper.cpp (AUR)."
exit 1
fi
if [ ! -r "$model" ]; then
notify-send -u critical "🎙️ dictate" \
"Model missing: $model. Install whisper.cpp-model-base (AUR)."
exit 1
fi
rm -f "$wav_file"
parecord --format=s16le --rate=16000 --channels=1 "$wav_file" \
>"$log_file" 2>&1 &
echo "$!" >"$pid_file"
notify-send -t 1500 "🎙️ Recording…" "Press the bind again to stop."
}
stop_and_transcribe() {
pid="$(cat "$pid_file")"
rm -f "$pid_file"
kill -TERM "$pid" 2>/dev/null || true
# Give parecord up to 2s to flush the WAV header.
i=0
while kill -0 "$pid" 2>/dev/null && [ "$i" -lt 20 ]; do
sleep 0.1
i=$((i + 1))
done
kill -KILL "$pid" 2>/dev/null || true
if [ ! -s "$wav_file" ]; then
notify-send -u low "🎙️ dictate" "No audio captured."
exit 1
fi
notify-send -t 1500 "🎙️ Transcribing…"
text="$(
whisper-cli -m "$model" -f "$wav_file" \
-l "$lang" -nt -np -t "$(nproc)" 2>"$log_file" |
tr -s '[:space:]' ' ' |
sed -e 's/^ //; s/ $//'
)"
if [ -z "$text" ]; then
notify-send -u low "🎙️ dictate" "Empty transcription. See $log_file."
exit 1
fi
printf '%s' "$text" | wl-copy
wtype -- "$text"
notify-send -t 2500 "🎙️ Dictated" "$text"
}
if is_recording; then
stop_and_transcribe
else
start_recording
fi
|
