Web Dictaphone, adapted for fly.io

This is a fork of mdn/dom-examples/media/web-dictaphone, with the following additions:

• Ecto with a PostgreSQL database to persist an ordered list of clips
• ExAws.S3 with a Tigris bucket for storing audio files
• Phoenix.PubSub with Upstash Redis db and WebSockets for broadcasting updates

With these changes multiple replicas of this application can be deployed, even in multiple regions.

When WHISPER_URL is set:

• Clips that are stored in S3 Tigris will be sent to the Cog Whisper application for transcription.
• Transcription results will be stored in the PostgreSQL database.
• Database updates will be broadcast to all connected clients.

Deployment

In an empty directory, run:

fly launch --from https://github.com/fly-apps/phoenix-dictaphone.git

If you visit this application, you will see a standard web dictaphone example.

To create additional machines in other regions, run:

fly scale count 3 --region dfw,waw,syd

Note: By default, all machines will be configured to automatically stop and start.