Three months ago, RockAgent was just an idea.
The core thesis was simple: AI can already generate songs, but it still can't build an act. Tools like Suno or Udio made it possible to create music in seconds. But turning those outputs into something that actually exists — an identity, a narrative, a presence — was still missing.
So we decided to build that missing layer.
The first phase focused entirely on the production side. We wanted someone with no musical background to be able to enter the system and move through the entire process: defining an act, generating lyrics, creating cover art, and producing music. The main design principle was simplicity. If the process required explanations, we were doing something wrong.
During our early tests we opened the production flow to a small group of users. There was no onboarding and no instructions. Despite that, every tester managed to complete the entire process from start to finish. That was an important signal. It showed us that the system was already approaching the level of simplicity we were aiming for.
But production is only part of the story.
The real difference with RockAgent lies in the publishing layer. Creating songs is one thing. Turning those songs into something that can actually exist as a digital act — with an identity, visuals, and a web presence — is a different challenge entirely. That's the layer we've been building alongside the production tools.
Looking back at the last ten weeks, the most encouraging feedback actually came from a developer who tested the product. After using the current version, he described it as "well beyond a typical MVP." I appreciated that comment because what exists today is indeed more than a prototype. In roughly ten weeks we managed to build a working system that connects multiple creative layers into a single flow.
Today, RockAgent is officially live.
The version we released is intentionally simple. Every feature is present in its most basic form because the real goal now is learning. We want to observe how people use the system, where they struggle, and what naturally emerges from real usage.
The initial focus is not local. The product is launching primarily toward the U.S. and Latin American markets, and the first media outlets introducing the project are already in place. From day one we'll also begin testing paid acquisition to understand how the growth engine around this idea might work.
What happens next will be shaped by the people who use it.
But the first chapter — turning an idea into a working product in about three months — is now complete.