Home

Voice Memos

I received great suggestions for taking voice memos and turning them into text. The two solutions:

Otter.ai Approach

Sign up for http://otter.ai/ - you get 600 minutes free. It's really high quality voice->text conversion.

gCloud Approach

  1. SETUP:
    • Enable Google Cloud Speech-to-Text API (think you have to set up billing)
    • Enable Google Cloud storage and create a bucket
    • Install gcloud command line tool
  2. Process the files

    for file in *.WMA; do ffmpeg -i "${file}" -c:a flac "${file/.WMA/.flac}"; done sox *.flac [file_name].flac

  3. Load into gcloud

    gsutil cp ./[file_name].flac gs://[bucket-name]/[file_name].flac

  4. Convert using API

    gcloud ml speech recognize-long-running gs://[bucket-name]/[file_name].flac --language-code="en-US" > [notes].txt

More involved then the otter approach, but if you have a lot of voice notes, over the free tier, might be cheaper.