Lyria 3

Lyria 3 (google/lyria-3-clip-preview and google/lyria-3-pro-preview, short names lyria and lyria-pro) generates instrumental music from a text description. There will be no vocals: these are beds for video, podcasts, backgrounds and idents. One request returns one track — unlike Suno, which returns two.

Tier	Short name	When to pick it
Pro	lyria-pro	When sound quality and arrangement matter more
Clip	lyria	Short inserts and backgrounds; cheaper

Parameters

A job is created with POST /v1/audio/music:

Field	Description
model	lyria or lyria-pro
prompt	Track description up to 5000 characters: genre, instruments, tempo, mood

mode, lyrics, style and title belong to Suno — Lyria ignores them.

Example

Python

import time

import requests

API = "https://api.mixen.ai/v1/audio/music"
headers = {"Authorization": f"Bearer {MIXEN_API_KEY}"}

job = requests.post(
    API,
    headers=headers,
    json={"model": "lyria-pro", "prompt": "Calm neoclassical piano with strings, 70 bpm, for a timelapse"},
).json()

while job["status"] in ("queued", "in_progress"):
    time.sleep(5)
    job = requests.get(f"{API}/{job['id']}", headers=headers).json()

mp3 = requests.get(f"{API}/{job['id']}/content", params={"index": 0}, headers=headers)
open("track.mp3", "wb").write(mp3.content)

cURL

curl https://api.mixen.ai/v1/audio/music \
  -H "Authorization: Bearer $MIXEN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "lyria-pro", "prompt": "Calm neoclassical piano with strings, 70 bpm"}'

Generation is asynchronous: create the job, poll GET /v1/audio/music/{id}, and once status: "completed" fetch the file from GET /v1/audio/music/{id}/content. The job lifecycle is covered in the Music guide; prices live in the catalog.