
Prompt clarity and structure
The workflow reads mood, era, and energy, then maps that to a full arrangement so sessions stay on brief.
Lyria 3 Pro is the full-song tier of Google's Lyria 3 music models, built for tracks with intros, verses, choruses and bridges. Describe your song here, then generate it with Lyria 3 Pro in the Lyria3 AI studio — an independent site, not a Google product.
Preview only on this page — generation happens in your dashboard.
Hear what the model can sketch before you refine. Clips load below the fold to keep the hero fast.
These clips show catchy lines: imagine drums, bass, and vocal doubles in a radio-friendly mix.




Pads and motion for focus: gentle dynamics so beds sit under dialogue.




Wide chords and rises for picture: balance impact without muddying dialog.




Three Lyria 3 tiers exist across Google's products and this studio. Here is how they differ before you pick one.
| What matters | Lyria 3 Pro | Lyria 3 | Lyria 3 Clip |
|---|---|---|---|
| Best for | Full songs with a clear arrangement | Quick song ideas and everyday tracks | Rapid drafts, loops and social-length assets |
| Track length according to Google | Up to about 3 minutes | About 30 seconds | About 30 seconds |
| Song-structure prompting | Yes — intros, verses, choruses and bridges can be prompted | Not documented by Google | Not documented by Google |
| Credits per generation on Lyria3 AI | Highest of the three — exact cost shown in the studio | Lower than Pro — exact cost shown in the studio | Lower than Pro — exact cost shown in the studio |
| How to select it in our studio | Pick "Lyria 3 Pro" in the studio model menu | The default model in the studio | Pick "Lyria 3 Clip" in the studio model menu |
For background on where Google offers each tier, see the Google Lyria 3 explainer
Clarity, speed, and repeatability — describe once and iterate without losing the vibe.

The workflow reads mood, era, and energy, then maps that to a full arrangement so sessions stay on brief.

Lead vocals stay present while supporting layers breathe; lows and highs translate to real speakers.

Grab WAV or MP3 and continue editing. Teams that ship weekly cut rework because outputs are predictable.
On the surface it is simple: describe, choose, generate — the same path every time.
Start from this page with a prompt, then land in the workspace. Your text and track type load automatically.
Adjust feel, lyrics, or instrumental mode so the next render matches your intent.
Create a take, listen, and iterate quickly to fit tight deadlines.
Creators who need cleared, high-fidelity music — short-form, podcasts, and ads.
Background tracks that match edits; skip long stock searches when you already know the mood.
Beds stay under voice with balanced dynamics so narration stays upfront.
Storyboard-aligned cues: iterations stay on-message during fast reviews.
Advanced sound with a simple path from idea to file — built for teams that ship.
Release-ready tone with fewer harsh artifacts so tracks sit in real mixes.
A full idea returns quickly so you can hit delivery slots without sacrificing taste.
Same steps each time: your team stays aligned from brief to bounce.
Real projects highlight speed and clarity — hear how teams put the tool to work.
Editors say cues land near final without endless tweaks.
Beds stay under voice and feel intentional, not generic.
Campaigns stay on brand while review cycles shrink.
Real projects highlight speed and clarity — hear how teams put the tool to work.
"Client scores stay on schedule; outputs feel balanced and ready to edit."
"Ad beds that used to take days now land in one session — huge for revisions."
"Our show intro gets compliments weekly; sounds custom, not stock."
"Client scores stay on schedule; outputs feel balanced and ready to edit."
"Ad beds that used to take days now land in one session — huge for revisions."
"Our show intro gets compliments weekly; sounds custom, not stock."
"I describe the scene and get a usable cue fast — fits my offline workflow."
"I prototype ideas before the studio so sessions stay productive."
"Stakeholders assume we hired a composer — quality holds up in review."
"I describe the scene and get a usable cue fast — fits my offline workflow."
"I prototype ideas before the studio so sessions stay productive."
"Stakeholders assume we hired a composer — quality holds up in review."
"Daily Shorts need fresh beds; unique textures without library fatigue."
"Game moods pivot hourly — quick iterations keep level teams unblocked."
"I hear poems as songs before committing arrangement time."
"Daily Shorts need fresh beds; unique textures without library fatigue."
"Game moods pivot hourly — quick iterations keep level teams unblocked."
"I hear poems as songs before committing arrangement time."
"Regional campaigns stay on brand without re-briefing every market."
"Vocal texture surprised us — great for pitch demos before final vocals."
"After fifteen years mixing, this is the first AI I trust in client folders."
"Regional campaigns stay on brand without re-briefing every market."
"Vocal texture surprised us — great for pitch demos before final vocals."
"After fifteen years mixing, this is the first AI I trust in client folders."
Access, sound, and workflow — the questions teams ask most.
Lyria 3 Pro is the full-song tier of Google's Lyria 3 family. On Lyria3 AI — an independent site, not a Google product — you select it in the studio model menu and generate structured songs with credits.
No. New accounts get a small number of welcome credits — currently five — and Lyria 3 Pro costs credits per generation; it is the most expensive of the three Lyria tiers in our studio. The exact cost is shown before you generate.
According to Google, Lyria 3 Pro generates tracks up to about three minutes and understands song structure — you can prompt intros, verses, choruses and bridges. Lyria 3 targets roughly 30-second tracks, and Lyria 3 Clip is optimized for fast, high-volume ~30-second clips. In our studio, Pro also costs more credits per generation.
Yes — pick instrumental in the hero and we carry it in the query so the studio opens with that toggle applied.
Yes — describe mood, range, and phrasing in your prompt for the best vocal placement.
Use the dashboard link from this page for history and export; your prompt comes with you.
Most jobs finish within a minute depending on queue; progress shows in the studio.
Common formats like MP3 and WAV for edit and delivery.
Many teams use it commercially; follow the account terms and license shown in the product.
Modern mobile browsers are supported; desktop offers more space for controls.
Add genre, tempo, mood, and instrumentation — specifics help the model match your brief faster.
Teams that need repeatable results fast without a long learning curve — that is where Lyria 3 Pro shines.
No. Lyria3 AI is an independent third-party website and is not affiliated with, endorsed by, or operated by Google or Google DeepMind. Google offers Lyria 3 Pro in the Gemini app, Google AI Studio, the Gemini API, Vertex AI and Google Vids.
Google documents Lyria 3 Pro tracks at up to about three minutes. The final length in our studio still depends on your prompt and settings.
One click from this page — your prompt and track type carry over so you can generate immediately.
Fields arrive pre-filled from this landing step.
Balances display in-app with the same rules as the rest of your account.
Support links live in the app whenever you need them.