The shift
Two things happened to music in the last two years.
First, AI generation crossed the listenability threshold. Suno and Udio shipped tracks people actually wanted on repeat. By 2026 a dozen well-funded providers can produce listenable music at low cost. Generation became a commodity input.
Second, TikTok finished what YouTube started. People stopped wanting to be passive listeners. They wanted to appear in the video. They wanted identity, not just taste.
Spotify cannot answer either shift — it is structurally a licensing business, locked into contracts with three major labels. Suno cannot answer either — it is structurally a creation tool, where the default surface is a prompt box. Both are stuck. The space between them is open.
What we're building
StarSinger is three surfaces, one platform.
A catalog of AI-generated music you can stream free, in a feed that behaves like a streaming app should — follow artists, save tracks, share URLs.
A creation studio with the lights on. Write a song with AI, paste a track from anywhere else, or upload your own. Clone your voice in 30 seconds and appear as the lip-synced star of the video. Watch every step the AI plans — stem separation, scene-by-scene shot list, lip-sync — and step in to tweak it. Most AI tools hide the pipeline; we expose it.
An open interface for AI agents. We were the first music platform to ship an MCP server, so Claude, Cursor, and the next wave of assistants can discover and recommend StarSinger music. As AI tools become primary discovery layers, music platforms without an open agent interface get cut out.
Why this isn't Spotify or Suno
Spotify is a licensing business dressed as a tech company. Its moat is exclusive contracts with Universal, Sony, and Warner. It cannot pivot to an AI-native catalog without destroying those contracts. Apple Music, YouTube Music, and Amazon Music are stuck the same way — bolting AI features onto a human catalog they can't disrupt.
Suno and Udio are creation tools. The default user journey starts with "what do you want to make?" — which filters out the 85% of music consumers who just want something good to play. Generation tools are the camera; we're trying to build the YouTube.
A handful of platforms are now competing to define AI-native music streaming as a category — Souna, BitzAudio, Pomelo.fm, and us. None has won. Each has a different bet.
Our bet
The platforms assuming listening is the endpoint are building Spotify with AI songs. The platforms assuming listening is the entry point to participation are building something genuinely new.
TikTok's split is roughly 85% consumers and 15% creators. The 15% is what makes the 85% sticky. Spotify has 0% creators inside the app — and the features its users actually share (Wrapped, playlists, decade rewinds) are the few places where they can express identity inside an otherwise passive platform.
In an AI-native world the barrier to creation collapses. You don't need to write, produce, play, or sing on key. You need a selfie and 30 seconds of voice. Every listener can, in principle, be an artist.
Listeners-become-stars is what makes the rest of the platform compound. The catalog gets sticky. The creator network grows itself. The agent layer has more to recommend. Three sides of the network feed each other.
Why open matters
We publish our pricing, our safety policy, and our technical approach. We ship an MCP server so agents can find our music. We make voice cloning a one-time $1.99 unlock instead of a recurring fee. We do this not because we are noble — we do it because the parts of the platform that should belong to creators belong to creators. Distribution, voice ownership, agent discovery, the right to opt out: not ours to gate.
We are also strict about consent. You can only clone a voice you have permission to use. We block non-consensual celebrity cloning, run moderation on lyrics and visuals, and remove violating content within hours.
Principles
- Consent is not negotiable. Voice cloning is opt-in, your voice, end of story.
- No subscriptions. Listening is free forever. Creation is pay-per-use. We'd rather customers actively choose each purchase than forget to cancel.
- Taste over volume. Anyone can generate thousands of AI songs in an afternoon. The work is choosing the few worth publishing.
- Openness. Public pricing, public safety policy, open MCP for agents, open distribution to the channels creators already use.
Who we are
StarSinger is a product of Veronata, Inc., a Delaware corporation based in San Francisco. We started in 2026. The team is small on purpose — engineers, producers, designers, and ML researchers who have spent careers building consumer products, music tools, and ML systems. We are backed by a mix of strategic and early-stage investors who share our view that consumer AI music will be a multi-decade category.
If any of this lands — as a creator, a partner, a candidate, or a curious skeptic — say hi at hello@starsinger.ai or via our contact form. Want to share Star Singer with your community? Our Refer & Earn program pays 20% on every purchase for a year.