Suno released v5.5 with Voices (personal voice cloning), Custom Models (train on your own catalog), and My Taste (auto-personalized style prompts).
Suno launched v5.5 of its AI music generation platform, introducing three major features: Voices (voice cloning via uploaded acapellas or direct mic recording with a verification phrase for anti-theft), Custom Models (train on a personal music catalog with a minimum of six tracks), and My Taste (automatic style inference from usage history). Voices and Custom Models are gated behind Pro and Premier subscriptions. The update marks a strategic pivot from fidelity improvements to user personalization and creative control.
Suno v5.5 is a product update, not an API release — developers can't directly access Voices or Custom Models programmatically yet. The voice verification anti-spoofing mechanism (requiring spoken phrase confirmation) is architecturally interesting but easily gameable with current voice synthesis models. If Suno opens an API for custom model inference, it would become a direct competitor to ElevenLabs for music-specific voice pipelines.
If you're building a music or audio app, test Suno's current API endpoints this week to benchmark output quality against Udio and document gaps — so you're ready to swap in Suno's custom model layer the moment they expose it via API.
Open your terminal and run: curl https://studio-api.suno.ai/api/generate/v2/ -H 'Authorization: Bearer YOUR_TOKEN' -d '{"prompt": "upbeat indie pop, female vocal, summer road trip"}'
Tags
Signals by role
Also today
Tools mentioned