Last updated: 2026-06-26
Google Flow Music by Google DeepMind creates full songs using Lyria 3 Pro. Plans from $6/mo with a free tier. iOS app live; formerly Riffusion/ProducerAI.
Google Flow Music is Google DeepMind's AI music creation platform, launched April 2026 as a rebrand of ProducerAI (formerly Riffusion, acquired by Google in February 2026). It uses Lyria 3 Pro to generate full songs up to 3 minutes at 44.1kHz stereo with a conversational Gemini agent for edits. Plans start free, with paid tiers from $6/month annual for 3,000 monthly credits. All outputs carry SynthID watermarks.
Google Flow Music is Google DeepMind's AI music creation platform, launched in April 2026 as the third identity of a product that started as Riffusion (an open-source project in December 2022), became ProducerAI in July 2025, and was fully acquired by Google in February 2026. The platform is powered by Lyria 3 Pro, a hierarchical attention model that generates complete songs up to three minutes long at 44.1kHz stereo quality. Every output is embedded with SynthID, Google DeepMind's imperceptible AI watermark, making it detectable as machine-generated content through the public SynthID Detector at synthid.google.com. The core experience is conversational: users interact with a Gemini-powered Producer agent in natural language to compose, arrange, and refine tracks. Instead of adjusting knobs or working on a piano roll, a user can say "make the bass punchier in the second verse" or "translate the lyrics to Spanish and keep the melody." Granular section editing, added in May 2026 at Google I/O, lets creators highlight any part of a track and change it without affecting the rest. A style transformation feature converts the genre or production style of full songs while preserving the original melody and structure, such as turning a pop track into a lo-fi study remix. Google Flow Music targets independent musicians, content creators, ad agencies, and developers. Music supervisors can generate custom, cleared tracks in minutes for sync licensing. Content creators get original, SynthID-watermarked music that sidesteps copyright claims. Developers can build browser-based audio plugins and custom DAWs directly on the Lyria 3 API through Vibe-code, a development workspace accessible via Google AI Studio. Pricing follows a credit-based model. The free tier exists with limited credits. Starter costs $6/month billed annually ($7.99/month monthly) for 3,000 credits per month, roughly 600 song generations. Plus costs $18/month annual ($24.99 monthly) for 10,000 credits. Member costs $48/month annual ($64 monthly) for 30,000 credits with up to 16 concurrent generations. Google AI Plus, Pro, and Ultra subscribers receive Flow Music credits as part of their Google AI subscription at Starter, Plus, and Member tiers respectively. The iOS app launched in May 2026; an Android app is in development. Google Flow Music integrates with the broader Google Flow creative suite alongside Flow (video generation) and uses Gemini Omni Flash for AI-generated music videos, added at Google I/O 2026. Music video creation is conversational: users direct the visual style, scenes, and pacing through natural language prompts that reference the track's narrative arc. Google is partnering with major music distributors including Believe to bring Flow Music to professional artists on their rosters.
Free tier with limited credits. Starter at $6/month (annual) or $7.99/month (monthly) for 3,000 credits (~600 songs). Plus at $18/month (annual) or $24.99/month (monthly) for 10,000 credits (~2,000 songs). Member at $48/month (annual) or $64/month (monthly) for 30,000 credits (~6,000 songs) with 16 concurrent generations. Google AI Plus, Pro, and Ultra subscribers receive Starter, Plus, and Member tier credits respectively as part of their Google AI subscription. Additional credits can be purchased as top-ups.
Google Flow Music is an AI music creation platform by Google DeepMind, available at flowmusic.app, that launched in April 2026. It is the third identity of a product that began as Riffusion in December 2022, an open-source hobby project by Seth Forsgren and Hayk Martiros that went viral for turning text prompts into music via Stable Diffusion spectrograms. The startup raised $4 million in seed funding in October 2023, rebranded as ProducerAI in July 2025 with a proprietary model called FUZZ, and was acquired by Google in February 2026 with the team joining both Google Labs and Google DeepMind. Google integrated Lyria 3 Pro, its most advanced music generation model, into the platform and relaunched it as Google Flow Music in April 2026, aligning it with the broader Google Flow creative suite. Believe, one of the world's largest independent music distribution companies, has partnered with Google to offer Flow Music to its roster of professional artists.
Google Flow Music offers a free tier with starting credits and four paid subscription tiers. Starter costs $6/month billed annually ($7.99/month billed monthly) and includes 3,000 credits per month, enough for roughly 600 song generations, with 2 concurrent generation slots. Plus costs $18/month annual ($24.99/month monthly) for 10,000 credits monthly, approximately 2,000 songs. Member costs $48/month annual ($64/month monthly) for 30,000 credits monthly, roughly 6,000 songs, with 16 concurrent generation slots and member-exclusive features. Credits do not roll over between billing cycles. Google AI subscribers receive bundled credits: AI Plus subscribers get Starter-tier credits, AI Pro subscribers get Plus-tier credits, and AI Ultra subscribers get Member-tier credits as part of their existing Google AI subscriptions. Additional credits can be purchased as top-ups at any tier.
Google Flow Music represents a complete generational leap from the original Riffusion. Riffusion generated short 30-second clips using Stable Diffusion applied to spectrograms, with no vocal capability and limited creative control. Google Flow Music uses Lyria 3 Pro, which generates full songs up to 3 minutes long at 44.1kHz stereo quality with vocals, lyrics, and full instrumentation. A conversational Gemini Producer agent replaces manual prompting, letting users direct changes in natural language. Granular section editing, added at Google I/O 2026, lets users change any section without affecting the rest of the track. AI music video generation via Gemini Omni Flash is entirely new. Style transformation converts a song's genre while preserving its melody. And all outputs carry SynthID, Google DeepMind's imperceptible AI watermark, which was absent from Riffusion's outputs.
Yes, Google Flow Music has a free tier that requires only a Google account to access, with no credit card needed. The free tier includes a starting allocation of credits for generating songs, but the number of free credits and monthly regeneration limits are smaller than paid tiers. Free-tier users have access to the core Lyria 3 Pro generation and the Gemini Producer agent for basic track creation and editing. Features like 16 concurrent generations, member-exclusive content, and the highest credit allowances require a paid plan. Google AI Plus, Pro, and Ultra subscribers receive bundled Flow Music credits as part of their existing Google AI subscriptions at Starter, Plus, and Member tiers respectively, making Flow Music essentially free-to-use for existing Google AI subscribers.
The two closest alternatives to Google Flow Music are Suno and Udio. Suno is the most widely used AI music platform with the largest community library of shared tracks and strong lyric generation, though it lacks granular section editing and its audio quality trails Lyria 3 Pro in complex arrangements. Udio signed licensing deals with Universal Music Group in October 2025 and Warner Music Group in November 2025, allowing it to generate music in the styles of real artists with artist permission, a capability Google Flow Music does not offer as of June 2026. For developers who need open-weight models, Stable Audio from Stability AI offers 6-minute tracks with fully licensed training data and open weights. AIVA is the strongest choice for classical composition with MIDI export, while Soundraw and Soundful serve brand music and professional sync licensing workflows.
Google Flow Music is best for independent musicians, songwriters, and content creators who want studio-quality original music without studio costs or music theory expertise. Content creators on YouTube, TikTok, and podcasts benefit most from SynthID-watermarked outputs that are clearly AI-generated and avoid copyright claims on platforms that use Content ID. Music supervisors and advertising agencies that need fast, original, cleared tracks for sync licensing can generate and refine full songs through the conversational agent in minutes. Developers building audio apps or browser-based DAWs can access Lyria 3 directly through Vibe-code and Google AI Studio. Google Flow Music is less suited to Android-only users (no Android app as of June 2026), enterprise teams needing Vertex AI API integration (not yet available for Gemini Omni Flash), or producers who require MIDI export and DAW integration.
Yes, developer access to Lyria 3 is available through multiple paths. The Gemini API on Google AI Studio supports Lyria 3 music generation for prototyping and development. Vibe-code, a developer workspace inside Google Flow Music itself, lets developers build browser-based audio plugins and custom DAWs directly on the Lyria 3 API using natural language, without requiring an audio engineering background. Lyria 3 and Lyria 3 Pro are also available on Google Cloud Vertex AI for enterprise deployments. As of June 2026, Gemini Omni Flash, the model used for AI music video generation within Flow Music, is available through the Gemini app and website but its Vertex AI API rollout was still in progress. Google does not currently support MCP (Model Context Protocol) connectors for Flow Music, though Lyria 3 is accessible via the Gemini API which is compatible with standard OpenAI-compatible client patterns.
Google Flow Music and Suno are the two most capable AI music platforms in 2026, with different strengths. Audio quality: Lyria 3 Pro's hierarchical attention mechanism maintains thematic consistency across full songs, which independent reviewers and music producers often rate above Suno's output for complex arrangements and orchestral elements. Suno leads on simple pop vocals and has a larger library of community-shared songs. Pricing: Google Flow Music's Starter plan at $6/month annual is cheaper than Suno's Pro at $8/month for comparable song lengths and quality. Editing: Google Flow Music's granular section editing lets users change any part of a track without affecting the rest, a capability Suno does not offer as of June 2026. Community: Suno has a significantly larger established user community and shared song library. Music videos: Flow Music adds Gemini Omni Flash for AI-generated music videos tied to tracks; Suno has no equivalent. Android app: Suno has an Android app; Flow Music does not yet.