Proof of Concept for transcoding podcasts into text using GCP Speech2Text service, following its NODE JS tutorial.
- Download this repo:
git clone https://github.com/emibcn/Podcast2Text.git
- Change directory into it:
cd Podcast2Text
- Create local directories:
mkdir flac credentials
- Create GCP credentials for consuming Speech2Text service at GCP IAM with -at least-
Service Usage Consumer
permission. - Copy credentials file to
./credentials
directory - Create
.env
file withGOOGLE_APPLICATION_CREDENTIALS=[CREDENTIALS FILENAME]
(without directory)
There is a script helper to transcode any audio file into text. It's syntax is:
./transcode.sh <FILEPATH> [START]
FILEPATH
: Path (relative or absolute) to podcast audio fileSTART
: Initial start seek (transcode beginning at this position). Same syntax as FFMPEG-ss
option.
This will encode the supplied file to FLAC format into ./flac
directory and then use the encoded file to send it to GCP Speech2Text service and get its transcription printed on screen.