The user had a finished meditation app. Macaly's AI agent generated four unique ambient tracks for it, one per meditation type. No stock audio, no sound libraries. One prompt, four moods.
One prompt. Four unique ambient tracks — one per meditation type. The agent didn't reach for files from an audio library or for sound samples. Every track was created purely from text, with nature sounds layered over ambient tones.
Generate music for my meditation app. About 3 minutes long. Only ambient and nature sounds, no singing.
The user asked for ambient music for each meditation. The agent familiarised itself with the meditation's content, picked a different mood for each one, and took care of generating the music under the hood. Four meditations, four moods, one prompt.
Every melody and every sound was created from a text prompt using AI music generation. No audio library, no samples. Each track is about three minutes long.
The meditation app was already finished. The agent didn't rebuild it, redesign it, or touch the rest of the project. It simply generated the four tracks and added them to the existing meditation cards. A new audio layer appeared without breaking anything in the app.
