Radio Starlight
Segment-based tool for designing AI-generated radio shows and podcasts from feeds, APIs, weather and music sources.
What it does
Radio Starlight is a program design tool for building generative radio stations and serialized podcasts. Creators assemble a show from segments, each bound to a data source such as RSS feeds, APIs, scraped web pages, weather data or music, with its own prompt, voice personality and background music. Generated shows and configurations can be published through an open StarlightCatalogService protocol for portability and remixing.
Core features
Segment-by-segment show builder
Multiple content sources per segment (feeds, APIs, pages, weather, music)
Cloud LLM or on-device Apple Foundation Model generation
AI-generated cover art and avatars
Open, publishable catalog format for shows and personalities
Custom personalities and background music sets
Best for
→Independent podcasters automating serialized episodes
→Radio broadcasters filling live blocks with AI-generated segments
→Studios building branded, consistent AI shows across sources
→Creators wanting to publish and share reusable show templates