Lyria 3 by Google Deepmind

This looks great! Would be nice to be able songs across smaller languages as well.

Google should consider creating a sheet music generator with complex harmonies and counterpoint.

music is a very interesting field. i was experimenting with elevenlabs and i am exited to see how googles compares to this.

The last thing the world needs right now.

Already thinking about how we can use this for CoreSight. Amazing achievement, congrats!

Has anyone gotten some good results with this they'd like to share? The generations feels pretty good, but still a bit slop-y.

I’m excited to hunt something that genuinely expands what “creative tools” mean in 2026.Meet Gemini with Lyria 3: Google DeepMind’s most advanced generative music model, now built directly into the Gemini app.This isn’t just another “type a prompt, get a beat” tool. It’s a full-stack AI music collaborator.How Lyria 3 works?Lyria 3 is a high-fidelity music generation model that lets anyone create 30-second custom tracks using:Text promptsPhotosVideosDescribe a vibe like: “A comical R&B slow jam about a sock finding their match” …and Gemini generates a polished, structured, catchy track, complete with lyrics and vocals.Upload a hiking photo of your dog? It composes a song inspired by it.This is music as expression, not production complexity.Core features and capabilities 1️⃣ Text → Full TrackGenre, mood, tempo controlLyrics auto-generated (or bring your own)Realistic, nuanced vocalsInstrumental-only options2️⃣ Image/Video → SoundtrackTurn visuals into high-fidelity musicMood-aware lyric writingPerfect for social content3️⃣ Creative ControlDirect vocal styleSpecify acoustic preferencesAdjust complexity and arrangement4️⃣ Professional-Grade OutputClean, structured tracksExport-ready audioCustom AI-generated cover art5️⃣ Built-In Safety & WatermarkingAll tracks embedded with SynthIDAI content verification inside GeminiWhy Lyria 3 is different?There are plenty of AI music tools out there. Here’s where this feels distinct:Musicality, not just loop generationNo lyrics requiredMulti-language supportVisual-to-Music pipelineResponsible AI LayerUse cases I am excited about:Personalized birthday songs in secondsInstant background music for Shorts/ReelsCommunity creators generating meme musicEmotional gifts (parents, partners, inside jokes)Indie creators prototyping soundtrack ideasMarketers testing vibe-based audio brandingThis lowers the barrier between idea → emotional artifact.My personal take: What’s powerful here isn’t that AI can “make music.” It’s that this removes friction between feeling something and expressing it musically.Over to you!Would you use this for content? For fun? Or as a serious creative tool? Let me know in the comments. :)