Helsinki-built. Producer-driven.
Songen launched in 2019 from Vibin Oy, a small mobile studio in Helsinki, Finland. The thesis was simple — most AI music tools chase the dream of "type a prompt, get a finished hit," which is fun for casual users but useless for working producers. Producers don't want finished tracks. They want raw material they can shape.
Six years later Songen is still in continuous development at the same studio, with the producer community as its primary collaborator. Most features in the current version started as user requests — the SharePlay support, the chord visualizer, the expanded MIDI export per-track separation, the genre engine updates that follow Billboard trends. The official product site lives at songen.app, with the iOS app on the App Store.
Honest trade-offs. Songen is iOS-only — Android producers can't run it natively, which is the single most common request in feedback channels. The model is instrumental-only, which means it doesn't compete with Suno or Udio on vocal generation. The on-device constraint means the model is smaller than what cloud services can run, so raw output complexity is more constrained than Suno's full-song-from-prompt output — but that's irrelevant if your workflow is "generate seed, edit heavily, finish in DAW." Pricing has shifted over the app's lifetime; check the App Store for current numbers before subscribing annually.
What Songen wins on, decisively: on-device generation (privacy, offline, speed), real MIDI export with per-track separation that drops cleanly into any DAW, the chord visualizer as a music theory learning tool while you compose, and the SharePlay collaboration feature that no major competitor has. For a producer who wants to integrate AI into a real workflow rather than replace it, this is the right architecture.