YOUR MUSIC, YOUR RULES
From an idea to playable music.
Instantly create an instrumental sketch on your phone, or use the genuine ACE-Step neural music model on a capable desktop to generate music with vocals. Each mode clearly explains how it works.
Original music arranger
Create a groove and melody.
PELIGROSA's own rule-based generator arranges percussion, bass, chords and melodic phrases. Play and download a WAV file. This mode is algorithmic synthesis, not neural AI, and does not generate vocals.
Choose a genre and click Create music.
Real open-source AI model
AI generation with vocals.
ACE-Step 1.5 is an open music model that can generate full music and vocals from your description. It runs in a supported desktop browser — no API keys or per-song API fees.
Before loading: model files are approximately 2 GB on first use. At least 4 GB of free RAM and a modern WebGPU browser are recommended. Typical phones may not be able to run this mode.
The open-source ACE-Step WebGPU project hosts the interface embedded below. Models are downloaded from external sources to your browser. PELIGROSA does not operate a GPU server.
The AI model won't load until you click the button.
ACE-Step is an independent open-source model; generated music is not Marco Moretto's official catalog. Check licenses, ownership of inputs and lyric rights before publishing.
Technology: PELIGROSA's original fast arranger uses synthesized sounds without external samples. Advanced mode integrates ACE-Step 1.5 through a community WebGPU interface. No paid AI API is required; external model file availability is outside PELIGROSA's control.