Describe it
Type it or speak it. “A slow, warm song about my grandmother’s kitchen. Acoustic. A little sad, but hopeful at the end.” That’s enough.
HUM turns plain words into real music. Describe the track you imagine — a mood, a story, a sound you love — and HUM composes, arranges and produces it. In the style of the artists that shaped you, or a sound that is entirely yours. No theory. No instruments. No barriers.
How it works
Type it or speak it. “A slow, warm song about my grandmother’s kitchen. Acoustic. A little sad, but hopeful at the end.” That’s enough.
Choose the style and features: the influences you love, tempo, mood, voice, instrumentation — or train HUM on your own past work so everything comes out sounding like you.
A finished, produced track in seconds — with stems, so you can keep shaping it. Iterate by talking: “more bass”, “sadder chorus”, “half the tempo”. It listens.
Why we exist
Making music today assumes years of training, expensive gear, and a body that cooperates: hands for the fretboard, sight for the timeline, fine motor control for the faders. Millions of people with music inside them are locked out — including many of the 1.3 billion people worldwide living with a disability. HUM is built accessibility-first, not accessibility-added: if you can express an idea, in any way you can express it, you can make real music.
If you can describe it, you can make it.
Music for every body. Every ability, every background, every idea — zero gatekeeping.
Value proposition
HUM collapses the distance between having a musical idea and holding a finished track — from years of practice to a single sentence.
The pitch
AI music generation crossed the quality threshold: full produced tracks from text are here. But today’s tools compete on model quality for the same prompt-savvy hobbyist.
The genuinely underserved market isn’t people who want better generation — it’s people who were never able to participate at all.
Accessibility-first creation. 1.3 billion people live with disabilities; music therapy, special education and care programs are established, paying markets that no music-AI company serves seriously.
Winning the hardest users produces the simplest product — which then serves the far larger “no musical training” mainstream.
Free to try; subscription for unlimited creation, personal style training and stem export.
B2B licensing to music-therapy clinics, schools and care organizations — recurring revenue with mission-aligned buyers, reachable without consumer-scale marketing budgets.
On style: HUM guides generation by influence and features — mood, era, instrumentation, production character — not by cloning any artist’s voice or likeness. Artist-consented style partnerships and licensed training data are core to the roadmap, because a tool built on inclusion can’t be built on taking.
Join us