The Scripts
TranscriptOMatic consists of the following scripts.
In addition: Don't forget to create the needed files and folders and set the right permissions.

Scripts in meetings/lib/
Scripts used by the executable scripts, providing basic configuration information.
Script: 
meetings/lib/paths.sh
Creates: 
Session Directory
Session Files 
audio.wav
transcript.txt
meta.env
Timestamps (ISO)
and
Provides path information for all scripts
#!/usr/bin/env bash
set -euo pipefail
BASE="$HOME/meetings/recordings"
mkdir -p "$BASE"
# ISO-like stamp with seconds
ISO_STAMP="$(date +%Y-%m-%dT%H%M%S)"
SESSION="$BASE/$ISO_STAMP"
i=0
while [[ -e "$SESSION" ]]; do
  i=$((i + 1))
  SESSION="$BASE/${ISO_STAMP}_$i"
done
mkdir -p "$SESSION"
AUDIO="$SESSION/audio.wav"
TRANSCRIPT="$SESSION/transcript.txt"
META="$SESSION/meta.env"
CURRENT="$BASE/.current"
Script: 
meetings/lib/whisper.sh
Contains:
the path to the local 
whisper.cpp installation
the language model used (currently 
ggml-tiny.en.bin) 
ASR (Automatic Speech Recognition) parameter used
#!/usr/bin/env bash
# whisper.cpp streaming binary
WHISPER="$HOME/whisper.cpp/build/bin/whisper-stream"
# Models
MODEL_EN="$HOME/whisper.cpp/models/ggml-tiny.en.bin"
MODEL_MULTI="$HOME/whisper.cpp/models/ggml-base.bin"
# Threading (Raspberry Pi 500 defaults)
INFER_THREADS=4
MEL_THREADS=16
# whisper.sh — Streaming / context parameters
STREAM_OPTS=(
  -c 1
  --keep-context
  --step 8000
  --length 12000
  --keep 800
)
# Thread options
THREAD_OPTS=(
  -t "$INFER_THREADS"
  -mt "$MEL_THREADS"
)

Script meetings/bin/meeting-start
Starts a new live transcription session.
./bin/meeting-start
Usage
German: 
meeting-start --de
English: 
meeting-start --en
Auto: 
meeting-start --auto
--de and --auto do NOT work at this stage of the project. Only use --en.
meeting-start --en
Script
#!/usr/bin/env bash
set -euo pipefail
HERE="$(cd "$(dirname "$0")" && pwd)"
source "$HERE/../lib/paths.sh"
source "$HERE/../lib/whisper.sh"
# ------------------------------------------------------------
# Argument parsing
# ------------------------------------------------------------
LANG_MODE="en"
NO_AUDIO=0
NO_TRANSCRIPT=0
while [[ $# -gt 0 ]]; do
  case "$1" in
    --de)            LANG_MODE="de";    shift ;;
    --en)            LANG_MODE="en";    shift ;;
    --auto)          LANG_MODE="auto";  shift ;;
    --no-audio)      NO_AUDIO=1;        shift ;;
    --no-transcript) NO_TRANSCRIPT=1;   shift ;;
    *)
      echo "Usage: meeting-start [--de|--en|--auto] [--no-audio] [--no-transcript]"
      exit 2
      ;;
  esac
done
if [[ "$NO_AUDIO" -eq 1 && "$NO_TRANSCRIPT" -eq 1 ]]; then
  echo "❌ --no-audio and --no-transcript together: nothing to do." >&2
  exit 2
fi
# ------------------------------------------------------------
# Select model + language options for whisper-stream
# ------------------------------------------------------------
case "$LANG_MODE" in
  en)
    MODEL="$MODEL_EN"
    LANG_OPTS=(-l en)
    ;;
  de)
    MODEL="$MODEL_MULTI"
    LANG_OPTS=(-l de)
    ;;
  auto)
    MODEL="$MODEL_MULTI"
    LANG_OPTS=()
    ;;
esac
# ------------------------------------------------------------
# Audio setup: persistent sink + virtual microphone
# ------------------------------------------------------------
DISCORD_SINK="discord_sink"
DISCORD_MONITOR="${DISCORD_SINK}.monitor"
VIRTUAL_MIC="whisper_mic"
echo "📁 Session: $SESSION"
echo "🗣️  Language mode: $LANG_MODE"
echo "🎙️  Audio recording: $([ "$NO_AUDIO" -eq 1 ] && echo 'off (--no-audio)' || echo 'on')"
echo "📝 Live transcript: $([ "$NO_TRANSCRIPT" -eq 1 ] && echo 'off (--no-transcript)' || echo 'on')"
echo "🔧 Setting up audio routing…"
echo "   Sink:   $DISCORD_SINK"
echo "   Source: $VIRTUAL_MIC (master: $DISCORD_MONITOR)"
echo "----"
# 1) Ensure discord_sink exists
if ! pactl list short sinks | awk '{print $2}' | grep -qx "$DISCORD_SINK"; then
  pactl load-module module-null-sink \
    sink_name="$DISCORD_SINK" \
    sink_properties=device.description=DiscordSink >/dev/null
fi
# 2) Try to move an active sink-input (Discord) to discord_sink
#    This may be transient in silent channels → poll briefly.
moved="no"
for _ in {1..40}; do
  SID="$(pactl list short sink-inputs 2>/dev/null | awk 'NF{print $1}' | head -n1 || true)"
  if [[ -n "${SID:-}" ]]; then
    if pactl move-sink-input "$SID" "$DISCORD_SINK" 2>/dev/null; then
      moved="yes"
      break
    fi
  fi
  sleep 0.25
done
if [[ "$moved" != "yes" ]]; then
  echo "⚠️  Could not move a sink-input to $DISCORD_SINK."
  echo "    Make sure Legcord is connected to a voice channel, then re-run meeting-start."
fi
# 3) Remove existing whisper_mic remap-sources (idempotent)
for mid in $(pactl list short modules | awk '$0 ~ /module-remap-source/ && $0 ~ /source_name=whisper_mic/ {print $1}'); do
  pactl unload-module "$mid" >/dev/null 2>&1 || true
done
# 4) Create virtual microphone (mono; attempt 16 kHz)
REMAPPED_MID="$(pactl load-module module-remap-source \
  master="$DISCORD_MONITOR" \
  source_name="$VIRTUAL_MIC" \
  channels=1 \
  rate=16000 \
  master_channel_map=front-left \
  channel_map=mono \
  source_properties=device.description=WhisperMic16k)"
echo "✅ Audio ready (remap module id: $REMAPPED_MID)"
echo "----"
# ------------------------------------------------------------
# Start recording (ffmpeg) and/or transcription (whisper-stream)
# ------------------------------------------------------------
FFMPEG_PID=""
WHISPER_PID=""
# --- ffmpeg: record to WAV ---
if [[ "$NO_AUDIO" -eq 0 ]]; then
  if [[ -e "$AUDIO" ]]; then
    echo "❌ Refusing to overwrite existing file: $AUDIO" >&2
    echo "   (Session dir collision or leftover file)" >&2
    exit 1
  fi
  ffmpeg -nostdin -y -hide_banner -loglevel error \
    -f pulse -i "$VIRTUAL_MIC" \
    -ac 1 -ar 16000 \
    "$AUDIO" &
  FFMPEG_PID=$!
  echo "🎙️  Recording to: $AUDIO"
fi
# --- whisper-stream: live transcription ---
# whisper-stream reads directly from the capture device (-c 1 in STREAM_OPTS).
# Do NOT pass -f: for whisper-stream, -f is a text output file, not an audio input.
if [[ "$NO_TRANSCRIPT" -eq 0 ]]; then
  "$WHISPER" \
    -m "$MODEL" \
    "${LANG_OPTS[@]}" \
    "${STREAM_OPTS[@]}" \
    "${THREAD_OPTS[@]}" \
    | tee -a "$TRANSCRIPT" &
  WHISPER_PID=$!
  echo "📝 Live transcription to: $TRANSCRIPT"
fi
echo "➡️  Follow the transcript with: meeting-follow"
echo "⏹  Stop with: meeting-stop"
echo "----"
# Announce active session
echo "$SESSION" > "$CURRENT"
# ------------------------------------------------------------
# Metadata
# ------------------------------------------------------------
cat > "$META" <<EOF
SESSION=$SESSION
AUDIO=${AUDIO:-}
TRANSCRIPT=${TRANSCRIPT:-}
LANG_MODE=$LANG_MODE
MODEL=$MODEL
NO_AUDIO=$NO_AUDIO
NO_TRANSCRIPT=$NO_TRANSCRIPT
DISCORD_SINK=$DISCORD_SINK
DISCORD_MONITOR=$DISCORD_MONITOR
VIRTUAL_MIC=$VIRTUAL_MIC
REMAPPED_MID=$REMAPPED_MID
FFMPEG_PID=${FFMPEG_PID:-}
WHISPER_PID=${WHISPER_PID:-}
EOF
[[ -n "${FFMPEG_PID:-}" ]]  && echo "$FFMPEG_PID"  > "$SESSION/ffmpeg.pid"
[[ -n "${WHISPER_PID:-}" ]] && echo "$WHISPER_PID" > "$SESSION/whisper.pid"
echo "✅ meeting-start finished setup successfully."

Script meetings/bin/meeting-stop
./bin/meeting-stop
Stops the active transcription session and finalises all session files.
Usage
meeting-stop
Script
#!/usr/bin/env bash
set -euo pipefail
BASE="$HOME/meetings/recordings"
CURRENT="$BASE/.current"
SESSION="${1:-$(ls -d "$HOME/meetings/recordings/"* | grep -v '^\.' | tail -n1)}"
META="$SESSION/meta.env"
# ------------------------------------------------------------
# Stop running processes
# ------------------------------------------------------------
if [[ -f "$SESSION/ffmpeg.pid" ]]; then
  kill "$(cat "$SESSION/ffmpeg.pid")" 2>/dev/null || true
fi
if [[ -f "$SESSION/whisper.pid" ]]; then
  kill "$(cat "$SESSION/whisper.pid")" 2>/dev/null || true
fi
# Remove current session pointer
rm -f "$CURRENT"
# ------------------------------------------------------------
# Load metadata (if available)
# ------------------------------------------------------------
if [[ -f "$META" ]]; then
  # shellcheck disable=SC1090
  source "$META"
fi
# ------------------------------------------------------------
# Clean up audio graph (optional but recommended)
# ------------------------------------------------------------
if [[ -n "${REMAPPED_MID:-}" ]]; then
  pactl unload-module "$REMAPPED_MID" >/dev/null 2>&1 || true
fi
# ------------------------------------------------------------
# Ask for meeting name
# ------------------------------------------------------------
read -r -p "Meeting name (for filenames): " MEETING_NAME
MEETING_NAME="${MEETING_NAME:-meeting}"
# Slugify: lowercase, umlauts, spaces → _
slugify() {
  local s="$1"
  s="$(printf '%s' "$s" | tr '[:upper:]' '[:lower:]')"
  s="$(printf '%s' "$s" | sed \
    -e 's/ä/ae/g' -e 's/ö/oe/g' -e 's/ü/ue/g' -e 's/ß/ss/g' \
    -e 's/[^a-z0-9 _-]/_/g' \
    -e 's/[[:space:]]\+/_/g' \
    -e 's/_\{2,\}/_/g' \
    -e 's/^_//' -e 's/_$//')"
  printf '%s' "$s"
}
SLUG="$(slugify "$MEETING_NAME")"
# ------------------------------------------------------------
# Final filenames
# ------------------------------------------------------------
ISO="${ISO_STAMP:-$(date +%Y-%m-%dT%H%M)}"
NEW_TRANSCRIPT=""
NEW_AUDIO=""
if [[ -f "$SESSION/transcript.txt" ]]; then
  NEW_TRANSCRIPT="$SESSION/${ISO}_${SLUG}_transcript.txt"
  mv -n "$SESSION/transcript.txt" "$NEW_TRANSCRIPT"
fi
if [[ -f "$SESSION/audio.wav" ]]; then
  NEW_AUDIO="$SESSION/${ISO}_${SLUG}_audio.wav"
  mv -n "$SESSION/audio.wav" "$NEW_AUDIO"
fi
# ------------------------------------------------------------
# Final output
# ------------------------------------------------------------
echo "⏹ Meeting stopped."
echo "📁 Session: $SESSION"
[[ -n "$NEW_AUDIO" ]]      && echo "🎧 Audio:      $NEW_AUDIO"
[[ -n "$NEW_TRANSCRIPT" ]] && echo "📄 Transcript: $NEW_TRANSCRIPT"
echo "🧾 Meta: $META"

Script meetings/bin/meeting-follow
./bin/meeting-follow
Display the transcript while it is generated.
Usage
meeting-follow
Script
#!/usr/bin/env bash
set -euo pipefail
BASE="$HOME/meetings/recordings"
CURRENT="$BASE/.current"
echo "📄 Waiting for active meeting session…"
# Wait until a current session is announced
while [[ ! -f "$CURRENT" ]]; do
  sleep 0.5
done
SESSION="$(cat "$CURRENT")"
META="$SESSION/meta.env"
TRANSCRIPT="$SESSION/transcript.txt"
# Check if live transcription is disabled for this session
if [[ -f "$META" ]]; then
  # shellcheck disable=SC1090
  source "$META"
fi
if [[ "${NO_TRANSCRIPT:-0}" -eq 1 ]]; then
  echo "⚠️  This session was started with --no-transcript."
  echo "   No live transcript available."
  exit 1
fi
echo "📄 Following transcript:"
echo "   Session: $SESSION"
echo "   File:    $TRANSCRIPT"
echo "----"
# Wait until transcript file exists
while [[ ! -f "$TRANSCRIPT" ]]; do
  sleep 0.5
done
tail -f "$TRANSCRIPT"
Expected behaviour:
meeting-follow will check the file 
$HOME/meetings/recordings/.current if an active transcript session exists. If not, it waits until the information about the active session has been added by 
meeting-start and will then start displaying the transcript.