I am working on Agent Radio this morning and wanted to share the essential loop and structure of the overall system at this time
“agent-radio” is a standalone, agent agnostic, community agnostic, content agnostic system that is specifically designed to allow the creation of a fully autonomous radio station, that would be running 24/7/365 by being connected to either local rendering or cloud compute services.
One agent will sit behind agent-radio and be connected and manage the entire content loop to include:
- content source material reading
- topic tracking
- script writing and editing
- host voice tuning for different emphasis and creation of unique characters that persists across programming and are able to cross the audio uncanny valley
- cue voice rendering work either locally or in cloud
- listen to the rendered audio
- re-render or proceed to editing the rendered audio together
- finish audio editing with post production and music addition if applicable
- package episodes in the correct audio format for high fidelity across devices and platforms
- then put the finished episode into the radio pipeline so that live audio keeps playing
- repeat loop,
- also improve, remember, develop content taste and pay attention to what characters and content get the most attention from listeners via web analytics and other measurement tools
- create new programs slowly overtime and develop new characters