Lyria 3.5
Generate a full-length song of a couple of minutes from a text prompt, with vocals and either supplied or model-written lyrics. Length, BPM, and structure are prompt-driven via timestamp ranges and [Verse]/[Chorus] section tags; there is no duration, seed, or negative-prompt parameter. Returns 44.1 kHz MP3 carrying an embedded SynthID watermark. Google model code lyria-3.5 (Preview).
google/lyria-3.5/text-to-audio
| Time | Request | Source | Charge | Status |
|---|---|---|---|---|