Pietro Passarelli

Podcast Transcription authoring tool

2020 Oct 10 WSJ

Stack: node, React, Javascript, bootstrap, react-bootstrap, firebase, STT, Google Cloud Functions,

An authoring tool to create transcriptions of podcasts, leveraging automated speech to text services.

An authoring tool to create transcriptions of podcasts, leveraging automated speech to text services.
  • Serverless architecture using Firebase.
  • It fetches new episodes via the RSS feed daily using a CRON job in a firebase cloud function.
  • It uses react transcript editor component to provide a UI to correct the automated transcription from STT services, while preserving word level time-codes.
An authoring tool to create transcriptions of podcasts, leveraging automated speech to text services.
An authoring tool to create transcriptions of podcasts, leveraging automated speech to text services.

BBC React transcript editor has got performance issues for transcript over one hour, but the average podcast for this use case is between 7 to 10 minutes, and max 20 minutes. So it's fine in this instance. But could replace with more recent slate-transcript-editor if requirements change.

Edit this page on GitHub