Google Lyria 3.5 AI Music Generator
Generate full-length songs with vocals using Google Lyria 3.5 on Epochal — timed lyrics, complete arrangements, and optional image-driven mood control.
Overview
Generate Complete Songs with Google Lyria 3.5
Google Lyria 3.5 turns a written description into a complete song — vocals, timed lyrics, and full arrangements from a single prompt. Describe the genre, mood, tempo, instruments, and song structure, or drop in a reference image to steer the atmosphere. One run produces a finished, ready-to-use track.
- Full-length songs with vocals
- Timed lyrics
- Text prompt + optional image input
- MP3 audio output
- Credits shown before each run

Explore audio models
More AI Audio Generators
Explore specialized AI audio models — each built for different creative styles and production needs.
Features
Google Lyria 3.5 Features

Prompt-to-Song with Vocals and Timed Lyrics
Describe a genre, mood, tempo, and vocal style, and Lyria 3.5 composes a full-length song — sung vocals aligned to timed lyrics and a complete arrangement, not a short loop. One prompt gets you a track you can drop into a video, podcast, or game scene.
Try a prompt
Image-Driven Mood Control
Upload a reference image and the model reads its atmosphere — palette, lighting, energy — to shape the theme of the music. A sunset photo steers warm and acoustic; a city-at-night shot steers electronic and moody. A way to direct sound without musical vocabulary.
Test image-to-music
Structure-Aware Composition
Prompt the song architecture directly: tell it where the verses, choruses, bridge, and outro go. The arrangement follows the structure you describe, which makes the output usable as real production music instead of a fixed-length loop.
Set your song structure
Browser-Based, Credits-Per-Run
Generate from your browser with no setup — the credit cost is shown before each run, and failed eligible runs are refunded automatically. Iterate on descriptions until the track fits, paying only for the runs you keep.
Generate a trackHow to use
How to Use Google Lyria 3.5
Describe the song you hear in your head, add an optional mood image, and generate a finished vocal track in one run.
Describe Your Song
Write what you want to hear — genre, mood, tempo, instruments, vocal style, and structure (verses, chorus, bridge). Lyria 3.5 reads the full description and composes a complete track around it.
Add a Mood Reference (Optional)
Upload one image to give the model a visual anchor for atmosphere and tone. A photo of a rainy street or a neon skyline nudges the arrangement toward the feeling you're after.
Generate and Listen
One click produces a full-length song with vocals and timed lyrics. Play it back right in the workbench and check whether the arrangement, vocals, and structure match your brief.
Refine and Re-run
Adjust the description — change the tempo, swap the instrumentation, or rewrite the mood — and generate again. Each run shows its credit cost before you commit.
Who should use
Who Should Use Google Lyria 3.5
Anyone who needs an original, finished song — from a description, in one run.
Video Creators & YouTubers
Need original background music and theme songs that no Content ID will claim. Describe the scene and mood — a full vocal track or instrumental arrives ready to drop under your edit.
Game & App Developers
Prototype soundtracks for menus, levels, and cutscenes without licensing rounds. Prompt the tempo and structure per scene, then iterate until the loop fits the gameplay.
Songwriters & Musicians
Sketch demo ideas fast: describe a verse-chorus structure, a genre, and a vocal style, and hear the concept as a complete arrangement before booking studio time.
Marketing & Ad Teams
Create custom jingles and campaign audio without stock-music libraries. A single prompt produces a vocal track matched to the brief, with the credits cost shown before each run.
FAQ
Google Lyria 3.5 FAQ
What it does, what a run costs, and when to reach for another model.
- What is Google Lyria 3.5?
- Lyria 3.5 is Google's AI music model. On Epochal, it turns a text description into a full-length song with vocals, timed lyrics, and complete arrangements. You can also add one reference image to guide the mood and theme of the track.
- Do I need to write lyrics?
- No. Unlike lyrics-driven models, Lyria 3.5 works from a single description — genre, mood, tempo, instruments, vocal style, and structure. It composes the vocals and timed lyrics from your prompt. If you want direct control over every lyric line, Mureka V9 takes paste-in lyrics instead.
- How much does a run cost?
- A generation costs 30 credits, shown before you run it. Failed eligible runs are refunded automatically, and paid plans add more credits.
- Is Google Lyria 3.5 free?
- No. Epochal runs on credits, and every generation shows its cost before you run it. New accounts get 20 credits after signing up, which is less than one Lyria 3.5 track (30 credits). Epochal is not affiliated with Google.
- What is the difference between Lyria 3.5 and Mureka V9?
- Both generate full songs. Lyria 3.5 composes everything — including the lyrics — from one prompt, with optional image mood control. Mureka V9 is lyrics-driven: you paste your own lyrics and get them sung. Choose Lyria 3.5 for prompt-only convenience; choose Mureka when the words must be yours.
- Can I upload Lyria 3.5 songs to Spotify or YouTube?
- The tracks are original MP3 files you can download and use in your projects. Streaming platforms don't reject music just for being AI-generated, but they do police artist impersonation and artificial streams — check the platform's current policy before you distribute, and note that Epochal is not affiliated with Google.
- How does Google Lyria 3.5 compare with Suno?
- Both turn prompts into full songs. Lyria 3.5 is Google's music model — it composes the vocals and timed lyrics from your description and accepts an optional reference image for mood. Suno is a separate subscription platform with its own catalog-style workflow. On Epochal you can run Lyria 3.5 pay-per-run instead of a monthly plan.
- Can I use the generated music commercially?
- Generated tracks are delivered as MP3 files you can download and use in your projects. Review the applicable terms for your use case, and note that Epochal is not affiliated with Google.
- What settings affect the output?
- Two inputs: your text description and an optional reference image. The more specific the description — genre, tempo, instrumentation, vocal style, and structure — the closer the track lands to what you intended.
Start now
Start creating with Google Lyria 3.5
Describe a song and generate it in your browser — no setup needed. The credit cost is shown before each run.
Try Google Lyria 3.5

