Why instrumental only?
Almost every creator who generates AI music wants a backing track they can talk over — vocals fight the voiceover and add IP risk. Lyria 3 Pro's prompt template explicitly asks for instrumental output (the genre presets all end with 'no vocals'). If you specifically want a vocal track, the AI Voice generator handles narration; the AI Music tool intentionally stays out of the singing-AI space.
Is the generated music royalty-free?
Yes. Google's Lyria license permits commercial use of the generated audio with no royalties owed. The track you download is yours to monetize on YouTube, TikTok, ads, podcasts — anywhere.
Why 30–90 seconds wall time?
Lyria 3 Pro is exposed via OpenRouter's chat-completions endpoint as an audio stream — the model takes a real chunk of compute time to generate each track end-to-end. ViralMint runs it on a worker Lambda with a 5-minute budget so even longer tracks complete reliably. The portal shows a live progress timer so you know where you are.
Can I generate variations of the same prompt?
Yes — every generation uses a fresh seed, so re-generating with the same prompt gives a different take. Most creators generate 3–5 variations of a prompt, keep the best, and discard the rest. At $0.15 each, that's well under a dollar to find the right vibe.
What if Lyria isn't available in the future?
ViralMint routes through OpenRouter, which gives us a single switch to swap in any other AI music model if Lyria pricing or availability changes. The desktop app's Smart Video pipeline also accepts your own uploaded music tracks as a fallback — see the Music Library page in the portal.