I asked Beatoven to make a minimalist investigative underscore and it understood what I meant

Quick question before I get into this: has anyone else tried describing music in fairly specific terms in Beatoven and been surprised by how well it interprets the brief?

I produce a true crime adjacent podcast and background music has always been a pain. Licensing anything decent is expensive and the free options all sound like the same five tracks everyone uses. Started using Beatoven a couple of months ago and the prompt-based generation is the thing I want to talk about.

I typed "minimalist investigative underscore, sparse piano, building tension, no resolution" and what came back was genuinely close to what I had in mind. Not perfect but close enough to edit from rather than reject entirely. The fact that it interprets qualitative descriptors like "no resolution" in a musically meaningful way rather than producing something random is what made me take it seriously.

The multiple model options make a real difference too. The Composer and Maestro engines produce noticeably different output from the same prompt. I run the same brief through both and pick the one that fits. For podcast work where the music needs to support rather than compete with the voiceover the difference between the two engines on a given brief is significant enough to matter.

The Video Sync tool is the one I use for any episode where there is a video version. You preview the generated track against your video in real-time and adjust before committing. Much faster than exporting and checking in a separate editor.