From Code to Chorus: How I Built a Virtual AI Artist
From Code to Chorus: How I Built a Virtual AI Artist
In September 2025, I started an experiment: could I create a virtual music artist using nothing but AI tools?
The answer turned out to be yes — and the results went further than I ever expected.
Meet AgnusBlast
AgnusBlast is a virtual AI artist I created using Suno, an AI music generation platform. The project spans multiple genres — K-pop, citypop, lo-fi chill pop, bubblegum pop, and ballads.
The creative process is collaborative: I write the concepts, moods, and lyrical themes, and Suno generates the compositions. Then I curate, iterate, and select the best versions. It's less "push a button, get a song" and more "direct an AI band through dozens of takes."
From Suno to Spotify
Creating the music was only half the journey. I wanted to see if AI-generated music could stand alongside human-created tracks on major platforms.
Using DistroKid, I distributed the tracks to:
- Spotify
- Apple Music
- YouTube Music
The distribution process is the same as any independent artist — upload the masters, set metadata, choose release dates, and DistroKid handles the rest.
The Breakthrough: Sugar Rush
Then something unexpected happened.
DistroKid has a community feature called Spotlight where artists submit songs and the community votes on them. I submitted "Sugar Rush" — a bubblegum K-pop track — and it was voted into the DistroKid Pop playlist.
That playlist has over 362,000 saves on Spotify. Sugar Rush was featured alongside tracks from other independent artists, exposed to hundreds of thousands of listeners.
The Artist Website
To give AgnusBlast a proper web presence, I built a promotional website using Lovable:
Try it live:
The site includes artist branding, track listings, and links to all streaming platforms.
The Full Stack
| Layer | Tool | Purpose |
|---|---|---|
| Music creation | Suno | AI composition and production |
| Distribution | DistroKid | Publishing to Spotify, Apple Music, YouTube |
| Artist website | Lovable | Promotional landing page |
| Cover art | AI image generation | Visual identity and album art |
What I Learned
- AI music is production-ready: The quality gap between AI-generated and human-produced pop music is smaller than most people think
- Curation > generation: The AI generates hundreds of variations — the real skill is selecting, iterating, and curating the best ones
- Distribution is democratized: DistroKid doesn't care if your music was made by a human, an AI, or a collaboration — the same pipeline works
- Community validation matters: Being voted into a 362K-save playlist by real listeners proved that AI-generated music can resonate with audiences
- The tools compound: Suno for music + DistroKid for distribution + Lovable for the website — each AI tool handles one piece of the puzzle
Is This the Future?
I think so. The AgnusBlast experiment showed me that the barrier to creating and distributing music has effectively collapsed. The remaining differentiator isn't technical skill — it's taste, curation, and creative direction.
The same pattern I see in AI coding applies to AI music: the human's role shifts from executor to director. You're not writing every note — you're shaping the vision and selecting the best output.