I built an AI music API on my own models to avoid the copyright mess
Hey IH
I'm Jina, solo founder of McRock.
Most AI music you can call via API runs through unofficial wrappers on models that are being litigated. As someone picky about genre who cared about shipping something legally clean, that didn't sit right — so I built my own engine.
McRock is context-driven (describe a vibe → it writes lyrics + composes) and genre-specialized for rock, post-punk, indie, hip-hop — instead of a generalist you re-prompt 20 times to hit your taste. Output is license-clean (our own models, commercial license), via an official REST API, a few cents per track.
I just opened the Enterprise API and I'm chasing first customers via build-in-public + outreach.
Two asks:
1. Honest feedback on the idea, output, and API design.
2. If you're building a game, app, or tool that needs music — reply and I'll send a free key (10 tracks) to try.
Happy to get into the build too (own model vs wrapper, the legal side, pricing).
What stood out to me is that you're competing on trust as much as generation quality.
A lot of AI music products focus on making better tracks. You're making it easier for businesses to actually ship those tracks without worrying about where they came from or whether they'll become a legal liability later. That's a very different value proposition than another music generator.