lyria 3 clip

Lyria 3 Clip (`lyria-3-clip-preview`) Model Card #

Overview #

Lyria 3 Clip (lyria-3-clip-preview) is Google's high-speed music generation model optimized for short clips, loops, social audio, and rapid style prototyping. It generates fixed 30-second stereo audio tracks along with synchronized lyric transcriptions and musical structure from text prompts or visual reference images.

Model Code: lyria-3-clip-preview
Release Date: March 2026
Status: Generative AI Preview
Documentation: Google Developer Music Generation ↗

Technical Specifications #

Property	Value
Input Modalities	Text, Image (up to 10 reference images)
Output Modalities	Audio (MP3 / WAV) and Text (Lyrics / Structure)
Input Token Limit	131,072 tokens
Audio Output Duration	Fixed 30 seconds
Audio Quality	44.1 kHz stereo, 192 kbps bitrate
Supported MIME Types	`audio/mp3`, `audio/wav`
Supported Languages	English, German, Spanish, French, Hindi, Japanese, Korean, Portuguese
Capabilities	Text-to-Music, Image-to-Music, Vocal generation, Instrumental mode, User lyrics
Watermarking	SynthID (Audio) + C2PA Content Credentials

Capabilities & Limitations #

Strengths #

Fast Prototyping: Low latency generation suited for iterating on musical genres, chord progressions, and tempos.
Multimodal Composition: Composes audio beds reflecting the palette, lighting, and mood of up to 10 input images.
Instrumental Isolation: Produces clean instrumental backing tracks without vocal artifacts when prompted with "Instrumental only".
Synchronized Lyrics: Returns lyric structure in the output_text / steps schema.

Limitations #

Fixed Duration: Always produces exactly 30 seconds. For longer compositions or multi-minute songs, use Lyria 3 Pro (lyria-3-pro-preview).
Single-Turn: Does not support sequential multi-turn audio inpainting or stem editing.

API Code Examples #

1. Python SDK (`google-genai`) - Interactions API #

python

import base64
from google import genai

client = genai.Client()

# Generate a 30-second chiptune clip
interaction = client.interactions.create(
    model="lyria-3-clip-preview",
    input="An energetic 8-bit chiptune arcade melody at 140 BPM in C major. Instrumental only, no vocals.",
)

if interaction.output_audio:
    with open("chiptune.mp3", "wb") as f:
        f.write(base64.b64decode(interaction.output_audio.data))

if interaction.output_text:
    print("Song Structure / Lyrics:\n", interaction.output_text)

2. JavaScript / TypeScript SDK (`@google/genai`) #

typescript

import { GoogleGenAI } from "@google/genai";
import * as fs from "node:fs";

const ai = new GoogleGenAI({});

async function generateClip() {
  const interaction = await ai.interactions.create({
    model: "lyria-3-clip-preview",
    input: "A 30-second lo-fi hip-hop loop with vinyl crackle and mellow electric piano at 85 BPM. Instrumental only.",
  });

  if (interaction.output_audio) {
    fs.writeFileSync("lofi_loop.mp3", Buffer.from(interaction.output_audio.data, "base64"));
  }
}

generateClip();

3. REST / cURL #

bash

curl -s -X POST \
  "https://generativelanguage.googleapis.com/v1beta/interactions" \
  -H "x-goog-api-key: $GEMINI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "lyria-3-clip-preview",
    "input": "A bright bossa nova guitar piece in F major. Instrumental only."
  }'

Lyria 3 Clip (lyria-3-clip-preview) Model Card #