Google Flow Music Complete Guide 2026: What It Is, How It Works & What Creators Can Do
Share
GOOGLE AI MUSIC · VERIFIED SEPTEMBER 10, 2026
Google Flow Music Complete Guide 2026: What It Is, How It Works & What Creators Can Do
Google Flow Music is no longer the only easy way to reach Lyria 3.5. Flow Music remains Google's dedicated creator-facing AI music workspace for generating, revising, organizing and developing music projects, while Lyria 3.5 is now also available directly in the Gemini app and Gemini API.
Lyria 3.5 is the music model. Flow Music is the dedicated music workspace. Gemini is now another direct access point. Google made Lyria 3.5 available globally in the Gemini app on September 4, 2026, with access also through the Gemini API, Google AI Studio and Google Vids. You do not need Flow Music just to generate with Lyria anymore; Flow Music matters when you want its deeper music-specific workflow.
Use Gemini when you want a fast general-user route into Lyria 3.5. Use Flow Music when you want the broader musician workflow around generation: references, targeted revisions, producer direction, projects, playlists, Spaces, sharing, downloads and music-video development. Google currently lets adults in supported regions sign up for Flow Music at no charge, with paid plans adding more credits, generations and model access.
Choose your Flow Music path
This page is the starting point. Use the focused guides below when you are ready to go deeper.
Flow Music vs Lyria vs Gemini: the names made simple
Google Flow Music
The dedicated music product/workspace. This is where the deeper song workflow, sessions, projects, playlists, editing, Spaces and music-video workflows live.
Lyria 3.5
Google DeepMind's current music-generation model. It creates music, lyrics and vocals and follows musical direction; it is the engine rather than a single standalone app.
Gemini
As of September 4, Gemini can generate directly with Lyria 3.5. In the Gemini app you can select or describe genre, choose vocal or instrumental output, use templates and choose shorter or longer tracks.
If you remember one sentence from this guide, make it this: Lyria is the engine; Flow Music is the dedicated music workspace; Gemini is now another direct way to use Lyria.
Do you need Flow Music to use Lyria 3.5?
No. This is the biggest change since the original version of this guide. Google says Lyria 3.5 is now available globally in the Gemini app and through the Gemini API. Google also lists access through Flow Music, Google AI Studio and Google Vids.
| Where you use Lyria | Best fit | JR takeaway |
|---|---|---|
| Gemini app | Fast general-user music generation | Start here when you want to describe a track, choose vocal/instrumental direction and generate without entering a larger music workspace. |
| Flow Music | Deeper music-specific creation and project workflow | Use it when revisions, references, organization, Spaces, sharing and connected music-video work matter. |
| Gemini API / Google AI Studio | Developer, prototyping and experimentation workflows | Useful when music generation needs to be integrated into a larger application or repeatable technical workflow. |
| Google Vids | Music inside a video/content workflow | Use when the music job is supporting a broader video project rather than managing an artist-song production workflow. |
The important decision is no longer “How do I get Lyria?” It is which Google surface matches the job I am trying to do?
What is Google Flow Music?
Flow Music is Google's dedicated AI music creation environment. It grew out of ProducerAI, which joined Google Labs in February 2026, and Google later folded the music experience into the wider Flow family. The product is explicitly built around more than generation: Google describes it as a place to create songs, playlists, music videos, Spaces and more.
That makes the useful comparison less “Google has a Suno clone” and more: Google is building a music workspace around generation, revision, organization, reusable producer behavior and the rest of its AI media ecosystem.
What is Lyria 3.5?
Google DeepMind published the Lyria 3.5 model card on July 29, 2026 and describes it as its newest music-generation model. DeepMind says Lyria can create tracks up to about three minutes, follow detailed musical direction, generate lyrics and vocals, respond to duration requests and use images as creative input.
For creators, the practical point is that Lyria is the engine, not the entire workflow. A model can generate strong audio; the surface you choose determines what you can do around that generation.
Who can use Flow Music, and is it free?
Google's current help documentation says you must be 18 or older, be in a supported region and sign up with a Google Account. Core Flow Music features can be used at no charge. More credits, generations and access to newer models come through Flow Music subscriptions or Google AI memberships.
| Google membership | Flow Music benefit mapping | Beginner takeaway |
|---|---|---|
| Free / no-charge signup | Core Flow Music features | Enough to learn the workflow before deciding whether more generation volume matters. |
| Google AI Plus | Flow Music Starter benefits | More capacity than the free level. |
| Google AI Pro | Flow Music Plus benefits | Higher-capacity creator use. |
| Google AI Ultra | Flow Music Member benefits | Top Google AI membership tier with the corresponding Flow Music benefit level. |
Google changes credits, model costs and plan details frequently. Check the current Flow Music account page before buying a plan based on a specific generation estimate.
What can you use to start a song?
Flow Music is not limited to a text box. Google's current song workflow supports text, image and audio references.
Text
Describe genre, mood, instrumentation, vocal style, lyrical subject, tempo and production direction.
Your audio
Add an existing audio file as creative reference material for the session.
Microphone recording
Record audio into the workflow so a rough sung idea, rhythmic phrase or other captured concept can become part of the creative conversation.
Images
Use an image as inspiration. DeepMind also documents image-to-music capability in the current Lyria experience.
How song creation actually works
A beginner can still start simply: describe the song and generate. The difference is what happens after the first result.
Give the producer a real brief
Describe the musical lane, emotion, instrumentation, vocal direction, subject and useful technical detail. A prompt such as “reggae” is a category. A brief explains what the song is supposed to do.
Generate and listen for decisions, not perfection
Judge the hook, voice, structure, groove, lyric meaning and sonic identity separately. Do not throw away a strong song because one transition is weak.
Talk to the producer
Google documents conversational edits such as shortening an intro, adding a bridge, making the kick more forceful, adding vocal reverb or changing instrumentation toward another style.
Target the section that failed
Use Replace, Extend or Cover rather than automatically regenerating the whole song. The dedicated editing guide shows when each tool makes sense.
Save the project logic
Keep sessions and songs organized inside projects, then group finished or related tracks into playlists when that helps.
Download only after you choose the version
Flow Music supports song downloads in selectable file formats. Keep the final exported file together with your lyrics, prompts, source inputs and release notes.
For the full song walkthrough, continue with How to Use Google Flow Music: Beginner Song Creation Workflow 2026.
Replace, Extend and Cover: why these matter
Replace
Select a section and ask Flow Music to change that part. Useful when the chorus works but the verse, transition or lyric passage does not.
Extend
Generate more music from a selected point. Useful for a longer ending, added section or continuation.
Cover
Describe a new version of the song. Think reinterpretation rather than merely “make another random generation.”
This is where Flow Music becomes more interesting than a simple prompt-to-song demo. The creator can protect the good decision and revise the weak one.
Compose controls and producer personalization
Flow Music exposes composition controls for areas such as Lyrics, Sound and Details, while its producer can also be shaped with persistent Instructions, reusable Google Flows, optional Memories, and producer roles including Instrument, Ghostwriter and Producer.
That does not mean a memory feature magically gives you a perfectly consistent artist identity. It means the product is trying to reduce how often you have to restate your working preferences.
Sessions, projects and playlists
Sessions
The working conversation around a song or creative task. Sessions can be renamed and shared.
Projects
Production containers that can hold sessions and the songs created in them.
Playlists
Collections of songs that can be organized and shared with the Flow Music community.
AI generation produces versions quickly. Without organization, creative abundance turns into version chaos.
Can you share, publish, remix and download songs?
Yes. Google's current song controls let creators share tracks, publish them publicly or keep them link-only/private, download them and control whether community remixing is allowed. Check those privacy and remix settings deliberately instead of accepting whatever default you happen to see.
What is available on mobile?
Google currently documents a dedicated Flow Music app for iPhone and iPad. The iOS workflow includes song generation, references, conversational editing, Replace/Extend/Cover, sharing, downloads, playlists and music-video creation.
Flow Music Spaces: where Google gets unusual
Spaces let creators describe and build custom musical tools instead of only generating songs. Google says Spaces can include musical instruments, games, experiences, plugins and other utilities. Spaces can be published, shared and remixed by the community and are currently documented as web/PC only.
See the full Flow Music Spaces guide for practical creator use cases.
Flow Music can also build music videos
Google Flow Music includes a dedicated music-video workflow. You can describe the video, add music, audio recordings or images as references, choose an available video model on desktop, then generate, publish, share and download the result.
A single idea can increasingly move through song → revision → project → artwork/reference → music video → publication inside connected Google AI tools. The AI Music Video Guide walks through that workflow in detail.
Does Google own the songs you make?
In its May 2026 Flow Music partnership announcement with Believe and TuneCore, Google stated that it does not claim ownership of original content generated with Flow Music.
That is useful—but it does not automatically answer copyright, commercial permission, distribution rules or third-party rights.
Platform ownership claim
Google saying it does not claim ownership tells you something about Google's position toward your generated output.
Copyright ownership
Whether a particular AI-assisted work qualifies for copyright protection, and which human-authored elements are protected, is a separate legal question.
For release decisions, read the dedicated Google Flow Music Rights & Commercial Use 2026 guide.
What about SynthID?
Google DeepMind currently says all Lyria tracks are imperceptibly watermarked with SynthID, allowing AI-created or AI-edited music to be detected. That is part of Google's broader provenance strategy for generated media.
Do not treat an invisible watermark as a substitute for your own documentation. Save your original inputs, prompt history, lyrics, human edits and release files anyway.
Flow Music vs Suno: where each approach feels different
If your only question is “Which one makes the better song?”, you are asking too narrow a question. The more useful comparison is which workflow helps you make the decisions you care about.
| Creator question | Flow Music direction | Suno direction |
|---|---|---|
| What is the core idea? | AI music workspace centered on a producer conversation, Lyria generation, revisions and connected Google media tools. | AI song-creation platform expanded into editing, identity, Studio and creator workflows. |
| Can I revise a section? | Yes: Replace, Extend, Cover and conversational edits. | Yes: section replacement, extension, Song Editor and Studio workflows. |
| Can my producer remember preferences? | Flow Music explicitly exposes Instructions, Flows and Memories. | Suno approaches repeatability through its own identities, models, editing and project tools. |
| Can I build custom music utilities? | Spaces are a distinctive part of Google's approach. | Not the same product concept. |
| Can I move toward AI video inside the same ecosystem? | Google has a direct Flow Music music-video workflow connected to its broader Flow/video models. | Suno's core strength remains the music-creation environment rather than Google's full media-model ecosystem. |
For the deeper generation-and-control comparison, read Google Lyria 3.5 vs Suno.
Who should try Flow Music first?
Try Flow Music if...
- You want to direct revisions conversationally.
- You want audio, image and text references in one creative environment.
- You care about section-level revision rather than endless full-song regeneration.
- You are interested in custom music tools/Spaces.
- You want music and AI video closer together.
- You already live in Google's AI ecosystem.
Start with Gemini if...
- You primarily want a fast path to Lyria 3.5 generation.
- You want genre, vocal/instrumental and duration choices without learning a larger music workspace first.
- Your immediate need is a backing track, jingle, personalized song or another quick generation job.
- You can move into Flow Music later if deeper music-specific workflow becomes useful.
There is no rule saying a serious creator must choose one Google surface—or one AI music platform—forever. The better question is which tool belongs at which stage of your process.
What I would try first as a beginner
- Decide whether your first job is quick generation or deeper production. If you only want to test Lyria, Gemini is now the simpler front door.
- Start free where available. Learn the workflow before buying generation volume.
- Choose one clear song idea. Do not test five genres at once.
- Write your own short creative brief. Subject, emotion, genre lane, tempo feel, instruments and vocal character.
- Generate two deliberate versions. Compare the same criteria: hook, voice, lyric, structure and production.
- If you need deeper revision, move into the Flow Music workflow. Protect what works and target the weak section.
- Save the chosen result and your notes. Start good version habits immediately.
- Download the chosen result and document your inputs and decisions.
Beginner FAQ
Is Google Flow Music the same as Google Flow?
No. They sit in the same broader Flow family, but Flow Music is the musician-focused environment. Google Flow itself is also a filmmaking product with its own video workflows.
Is Lyria 3.5 a separate app?
No. Lyria 3.5 is Google's music-generation model. As of September 2026, Google makes it available through multiple surfaces, including the Gemini app, Gemini API, Flow Music, Google AI Studio and Google Vids.
Do I need Flow Music to use Lyria 3.5?
No. Lyria 3.5 is now available directly in Gemini. Use Flow Music when its deeper music-specific creation, revision and project workflow is the reason you are there.
Can Lyria make full songs with vocals and lyrics?
Yes. DeepMind documents lyrics, vocals and cohesive tracks up to about three minutes for the current Lyria experience.
Can I upload my own audio in Flow Music?
Yes. Google's current Flow Music song workflow supports audio files and microphone recordings as references.
Can I change only one part of a Flow Music song?
Yes. Replace targets a selected section; Extend continues from a point; Cover creates a new interpretation.
Can I download my songs?
Yes. Flow Music includes song downloads and lets you select a file format.
Is Flow Music free?
Google says all core Flow Music features can be used at no charge. Paid Flow Music or Google AI plans provide more credits, generations and model access.
Can I commercially release what I make?
Google's ownership language, your human authorship, third-party inputs, copyright law, distributor requirements and platform rules are separate issues. Read the dedicated rights and commercial-use guide before treating a generated track as release-ready.
GOOGLE FLOW MUSIC 2026 SERIES
Start with the job, then choose the Google surface.
If you only need a fast Lyria generation, Gemini is now a valid front door. If you want a deeper music-production workflow, continue into Flow Music. If you already have a song, jump to editing. If you are preparing a release, go straight to rights and commercial use.
Start the Flow Music WorkflowCheck Rights & Commercial UsePrimary Google sources
- Google — Create your best tracks yet with Lyria 3.5 in Gemini (September 4, 2026)
- Google Flow Help — Get started with Google Flow Music
- Google Flow Help — Create & manage songs in Flow Music
- Google Flow Help — Organize sessions and projects
- Google Flow Help — Playlists
- Google Flow Help — Spaces
- Google Flow Help — Music videos
- Google Flow Help — Credits and subscription plans
- Google DeepMind — Lyria 3.5
- Google DeepMind — Lyria 3.5 model card
- Google — Believe / TuneCore / Flow Music partnership
Last verified: September 10, 2026. Product capabilities, credits, supported regions and model access can change quickly. Rights discussion is creator education, not legal advice.
