Transcribe a WhatsApp voice note

You will build: a command that transcribes every voice note in a folder and writes the text beside each one. You will need: curl or Python with the Python library (pip install namu-ai), and NAMU_API_KEY set.

Get the voice note onto your computer

In WhatsApp, long-press the voice note → Share (Android) or Forward → Share (iPhone) → send it to yourself by email or save it to Files. You get a file like PTT-20261108-WA0003.opus, which is Ogg Opus. Keep it as it is: the API reads it directly.

Transcribe one

curl https://api.namuai-lab.com/v1/transcribe \
  -H "Authorization: Bearer $NAMU_API_KEY" \
  -F file=@PTT-20261108-WA0003.opus \
  -F language=ha

Send dje for Zarma or fr for French.

Transcribe a whole folder

transcribe_folder.py

import sys
from pathlib import Path

from namu import Namu, NamuError

client = Namu()
folder, language = Path(sys.argv[1]), (sys.argv[2] if len(sys.argv) > 2 else "ha")
AUDIO = {".opus", ".ogg", ".m4a", ".mp3", ".wav", ".aac", ".webm", ".flac"}

for note in sorted(p for p in folder.iterdir() if p.suffix.lower() in AUDIO):
    text_file = note.with_suffix(".txt")
    if text_file.exists():
        continue
    try:
        result = client.transcribe(note, language=language)
    except NamuError as e:
        print(f"{note.name}: {e.type}")
        continue
    text_file.write_text(result["text"] + "\n", encoding="utf-8")
    print(f"{note.name} ({result['audio_seconds']:.0f} s): {result['text'][:60]}")

python transcribe_folder.py ./voice-notes ha

Good to know

The text comes in lower case, without punctuation.
Notes longer than 5 minutes return audio_too_long: cut them first.
Ask before you transcribe someone’s voice note: it is their voice and their words.

Build an interpreter

Translate what was said

Speech to text

Every option

​Good to know

​Next

Build an interpreter

Speech to text

Good to know

Next