Amped AI will shut down on 28 February 2026. Read the full message below.
Live 24/7 100% AI Radio

Amped AI — Always on. Always different.

A new kind of station: AI-generated music, shows, and voices. Streaming non-stop with real programming, real hosts, and zero humans behind the mic.

Car mode enlarges the player controls, hides secondary content for easier in-car use, and may immediately resume playback when enabled.
View Schedule Recently Played
On Air

AMP
Up Next
  1. Loading schedule…

Amped AI will be shutting down on 28 February 2026.

This project started as an experiment, a way to explore what fully AI-generated radio could look like, how far the technology could be pushed, and where it clearly still falls short. Along the way, it proved something important: AI can create atmosphere, consistency, and even moments of genuine connection, but it also highlighted the value of intent, context, and human judgment.

Amped AI was never meant to replace people. It was meant to test ideas, question assumptions, and see what happens when you let a system run end-to-end. In that sense, it did exactly what it was supposed to do.

Thank you to everyone who tuned in, listened, experimented alongside it, or simply let it play in the background.

The signal goes quiet on 28/02/2026.

- Marc

Featured Shows

Schedule

Recently Played

    About Amped AI

    Amped AI on-air hosts

    Amped AI is a fully automated radio station running around the clock. Shows are written by large language models, voiced using advanced text-to-speech, and scheduled with AzuraCast. Every track played is AI-generated music, making the station a complete end-to-end experiment in machine-driven broadcasting.

    The goal is simple: see how close AI can get to creating the energy and flow of traditional radio, while opening the door to new kinds of programming that wouldn’t be possible with a fixed human schedule.

    Live 24/7 Powered by AzuraCast AI Hosts 100% AI Music

    Ethics & Transparency

    Amped AI is an experiment — a side project, not a full-time job for anyone involved. If people enjoy it, great; if not, that’s fine too. The goal isn’t to replace radio DJs or take work away. Every show you hear is fully generated using a mix of large language models and ElevenLabs text-to-speech, with light human curation to keep things sensible and safe.