Lyria 3 Pro API
by Google
Full songs from text or images, in studio-grade stereo.
Google DeepMind's Lyria 3 Pro music model (early 2026) through one unified API. Full songs up to around three minutes in 44.1 kHz stereo, generated from text prompts or reference images.
- ~3 min
- Full songs
- 44.1kHz
- Stereo audio
- 10
- Reference images
- 10 credits
- Flat per track
Playground
Try Lyria 3 Pro right here.
Capabilities
What's new in Lyria 3 Pro.
Full songs, not loops
Lyria 3 Pro understands song structure: it composes tracks up to roughly three minutes with distinct verses, choruses, and bridges that stay musically consistent from first note to last.
44.1 kHz stereo fidelity
Output is high-fidelity 44.1 kHz stereo, DeepMind's biggest jump in audio quality over Lyria 2, with cleaner mixes and clearer vocals.
Image-to-music
Attach up to 10 reference images and the model composes music inspired by their mood, colors, and atmosphere, a soundtrack for a scene you show it rather than describe.
Time-aligned lyrics
Outline the progression of a song in your prompt and control when lyrics start and end. Vocals, user-provided lyrics, and instrumental mode are all supported.
Tempo and structure control
Condition the tempo, set duration and intensity, and shape the arrangement in natural language, granular control without a DAW.
Watermarked and provenance-ready
Every track carries the SynthID watermark and C2PA Content Credentials, which matters for platforms that need to label AI-generated audio.
Showcase
See what Lyria 3 Pro can create.
Real outputs generated with Lyria 3 Pro on Apiframe, each with the prompt behind it.
Score this coastal dawn harbor still: warm indie-folk, female vocals, fingerpicked guitar, leaving town at first light.
A hushed harbor-night song, brushed drums, upright bass, female vocal, foghorns far away, indie folk.
A glasshouse morning song, light piano, soft choir pads, hopeful, indie-folk instrumental, around two minutes.
A warm indie-folk song about leaving a small coastal town at dawn, female vocals, fingerpicked acoustic guitar, soft strings building in the final chorus, around two minutes.
Pricing
Simple per-generation pricing.
10 credits per generation ≈ $0.10
- You only pay for successful generations. Failed jobs are refunded automatically.
- One credit balance across every model on Apiframe, no per-model plans.
- Start with free credits. No subscription required.
- Volume discounts on larger plans.
Specs
At a glance.
Comparison
Lyria 3 Pro vs Lyria 3 Clip vs Udio
Google's flagship against its speed tier and the vocal-realism leader. One key, three options.
| Feature | Lyria 3 Pro | Lyria 3 Clip | Udio |
|---|---|---|---|
| Max length | ~3 minutes | 30 seconds | 130 seconds |
| Signature strength | Song structure + image-to-music | Fast loops and stingers | Vocal realism and deep controls |
| Avg. generation | ~90s | ~20s | ~90s |
| Price per track | 10 credits flat (≈$0.10) | 5 credits flat (≈$0.05) | 9–35 credits (≈$0.09–$0.35) |
| Best for | Finished songs and scored scenes | Prototyping and social clips | Genre work where the vocal carries it |
Specs reflect what each model supports on Apiframe today, from the live catalog.
Quickstart
How to call the Lyria 3 Pro API.
Send one POST /v2/music/generate request with your API key. The call returns a jobId you can poll, or pass a webhook_url and the result is pushed to you the moment it's ready.
Swap lyria-3-pro for any other model and nothing else changes.
const response = await fetch("https://api.apiframe.ai/v2/music/generate", {
method: "POST",
headers: {
"X-API-Key": "afk_your_api_key_here",
"Content-Type": "application/json",
},
body: JSON.stringify({
"prompt": "an upbeat synthwave track with a driving bassline",
"model": "lyria-3-pro",
"lyriaParams": {
"reference_images": [
"https://example.com/input.jpg"
]
}
}),
});
const { jobId } = await response.json();
console.log(jobId); Input schema
Every field Lyria 3 Pro accepts, with types and defaults. The full reference lives in the docs.
| Field | Type | Description |
|---|---|---|
| prompt required | string | Text description of what to generate. |
| model required | string | The model identifier for this endpoint. Default:"lyria-3-pro" · "lyria-3-pro" |
| lyriaParams.reference_images | string[] (URLs) | Optional moodboard — Lyria uses these to inspire the composition. |
Lyria versions
Every Lyria music model on Apiframe. Switch with one parameter.
Use cases
What teams build with Lyria 3 Pro.
Soundtracks from a scene
Pass up to 10 images from a video, game, or brand world and Lyria composes music that matches their mood, scoring by showing, not describing.
Complete songs for content
Verse-chorus-bridge structure over up to three minutes means finished songs for videos and campaigns, not loops that need arranging.
Lyric-driven briefs
Time-aligned lyrics let you outline when each line lands, write the hook, place it in the chorus, and the arrangement follows.
Platforms that need provenance
SynthID watermarking and C2PA Content Credentials are built into every track, which keeps AI-labeling policies simple to satisfy.
Reviews
9,800+ developers. One API.
Real Apiframe reviews from Trustpilot and G2.
Honestly the best AI API
They got every engine possible and it's great to have them all in one place.
Excellent variety, low latency
The variety of models is excellent and the latency is low.
One simple API for powerful AI models
The best thing is that it gives access to many strong AI models in one simple API. Before, I had to rely on separate platforms for each task, but now I can handle everything through one API.
Lyria 3 Pro questions
What teams ask before shipping with Lyria 3 Pro. Everything else lives in the docs.
What is Lyria 3 Pro?
Google DeepMind's premier music generation model (early 2026). It composes full-length songs up to about three minutes, with vocals, structure, and full instrumental arrangements, in 44.1 kHz stereo.
How is it different from Lyria 3 Clip?
Pro is the full-song tier with professional structural awareness: verses, choruses, and bridges over ~3 minutes. Clip is optimized for speed and generates 30-second loops and stingers.
Can it really generate music from images?
Yes. You can pass up to 10 reference images alongside the text prompt, and the model composes a track inspired by their mood, style, and atmosphere.
Does it support lyrics and vocals?
Both. It generates vocals with its own lyrics, accepts lyrics you provide, supports time-aligned lyric placement, and has a pure instrumental mode.
How much does it cost?
A flat 10 credits (≈$0.10) per generation on Apiframe, with a typical completion time around 90 seconds.
Where can you access it?
Through Apiframe with one API key, async jobs, and webhooks, alongside every other music, image, and video model in the catalog.
Related models
All music modelsSuno
by Suno
Full songs with lyrics, vocals, and instrumentation.
View modelUdio
by Udio
High-quality music generation with custom lyrics.
View modelProducer
by Producer
Producer-grade music with seed and lyrics control.
View modelElevenLabs Music
by ElevenLabs
Long-form composition (up to 5 min).
View model9,800+
developers & companies
Official partner of
Power your next AI product with Apiframe.
Instant access to 70+ media models through a single API. Start free and scale when you're ready.