Skip to content
C-Kernel-Engine docs Antshiv Open Source

C-KERNEL-ENGINE · GUIDE

Transcribe a recording

Turn an English WAV into text

Start with a short recording whose words you know. With FFmpeg and jq installed, replace recording.wav with your input file and run from the CKE checkout.

Examplebash
mkdir -p "$HOME/cke-audio/results"
ffmpeg -i recording.wav -ac 1 -ar 16000 -c:a pcm_s16le "$HOME/cke-audio/input-16k.wav"
RUN_DIR="$HOME/.cache/ck-engine-v8/models/whisper-base-local"

version/v8/scripts/cks-v8-run audio hf://openai/whisper-base \
  --run "$RUN_DIR" \
  --wav "$HOME/cke-audio/input-16k.wav" \
  --language en \
  --task transcribe \
  --max-tokens 448 \
  --output "$HOME/cke-audio/results/transcript.json"

jq -r '.decoder.transcript_text' \
  "$HOME/cke-audio/results/transcript.json" \
  > "$HOME/cke-audio/results/transcript.txt"

Open $HOME/cke-audio/results/transcript.txt and compare it with the recording. Check proper names and numbers by ear.

Which recognizer?

Whisper Tiny, Base and Small have public short-fixture results. Base also has a 33-second comparison. Parakeet TDT 0.6B v3 has an English native transcription path. Cohere Transcribe 2B accepts externally supplied speech segments for long recordings. Results from earlier 42-minute reference runs are separate from the native-host evidence. Parakeet instructions · Cohere instructions