Lyria
Shape the future of tech through early experiments.
Product Demo Video
Screenshots
Distribution Score: 40/100 What is this? ⓘ
What is Lyria?
Developed by Google DeepMind, Lyria 3 is a highly sophisticated AI model specifically designed for music generation. This model offers a creative platform for users to explore, experiment with new ideas, and create their desired sound.
Users can upload an image and Lyria 3 has the ability to transform it into a custom high-fidelity track. Lyria 3 allows for unique customization, as it provides the ability to define realistic vocal styles and acoustic preferences, thereby enabling users to produce music that aligns perfectly with their creative vision.
The output can range from background ambience to mainstage anthems, designed to produce crisp and clear tracks ready for various projects. Additionally, Lyria 3 fits into Google's mission of building AI responsibly to benefit humanity, and prioritizes AI safety through proactive security, untouched by evolving threats.
Lyria 3 Pro, released March 2026, extends the model's generation capability to tracks up to three minutes in length with professional-grade structural awarenessunderstanding verse-chorus-bridge architecture, intro and outro conventions, and the energy dynamics that characterize different music formats.
Earlier AI music generation was constrained to short loops or atmospheric clips; Lyria 3 Pro produces tracks with the compositional structure of professionally produced music, including tempo conditioning for precise BPM targeting and time-aligned lyrics that follow a user-defined progression.
SynthID watermarking embeds an imperceptible signal in every Lyria-generated track that can be detected by Google's identification system, establishing provenance for AI-created music.
This technical approach to attribution addresses the attribution challenge that has complicated AI music deployment in professional contexts: creators using Lyria outputs can demonstrate the tracks were AI-generated rather than recorded samples of existing artists, and the watermark persists through standard audio processing and format conversion.
Lyria 3's integration across Google's product suiteavailable in the Gemini app for conversational music generation, in Google Vids for marketing and presentation soundtrack creation, and through the YouTube Dream Track program for creator contentpositions it as infrastructure rather than a standalone tool.
Who is Lyria for?
Learn this stack in Academy
Get implementation playbooks for tools like Lyria in guided Academy lessons. Start free, then unlock the full library with Learner.
Open Academy →Release History
Longer track generation, expanding from shorter creations to tracks up to 3 minutes. Better structural awareness, with explicit control over sections like intros, verses, choruses, and bridges. Stronger customization and creative control for more complex song builds and transitions. Broader rollout across Google products, including Vertex AI, AI Studio, Gemini app, Google Vids, and ProducerAI. More production-oriented use, pushing it beyond creative sketching into scalable soundtrack, video, and song-making workflows. 2 0
Cohesive track generation with smoother musical flow from start to finish. Supports vocals and can generate songs with lyrics from prompts. Multimodal prompting that can turn an uploaded image into a matching track. Broader genre and language coverage for vocals and styles. Ships inside Gemini as a music-making feature, not just a developer-only or lab-only demo. 4 0
Initial release of Lyria.
Pricing & Access
Pricing details on provider page.
Alternatives to
Frequently Asked Questions
Founder
Comments (0)
Sign in to join the discussion.