Voice as a compositional medium.

Voice AI had a commodification problem, and I watched it happen from the inside. Working as a lead intelligence designer in enterprise environments, I saw organizations reach for the same shelf of synthetic voices, uniform, unlicensed, artistically hollow, and call it a solution. The artists whose performances trained those models saw nothing. The teams deploying them had no meaningful way to shape what they were building. The tool and the ethics were both broken.

That observation became Lyric. I cofounded the platform not as a technical exercise but as a response to a design failure: the absence of a system that treated voice as a compositional medium and artists as partners rather than source material. The foundation we built was intentional. Every voice on Lyric was performed by a real artist who shaped it, named it, and earns from every deployment. That ethical structure was not separate from the product. It was the product's premise.

The composer followed from that premise. Each voice in the edition was built for a specific use case, not pulled from a catalog of thousands. Inside the composer, the user directs the voice the way a producer would in a studio session: shaping the read, adjusting the performance, refining until the output fits the work. The result is audio that sounds like it was made for the brand, because the direction came from the brand.

The ethics are not a layer on top of the product. They are the product. A real artist performed the voice. A real user directs it. That chain is what makes the output worth anything.