Unified music generation framework
ByteDance Seed-Music combines auto-regressive language modeling and diffusion in one framework that adapts to evolving musician workflows.
Powered by ByteDance Doubao Seed team's Seed-Music suite. Seedmusic supports score-to-song conversion, controllable generation, music and lyrics editing, and low-threshold voice cloning at 44.1kHz stereo quality.
Start Free NowByteDance Seed-Music combines auto-regressive language modeling and diffusion in one framework that adapts to evolving musician workflows.
Convert musical scores into full vocal and instrumental tracks with interpretable lead sheet tokenization for precise control.
Condition on natural language style descriptions, audio references, musical scores, and voice prompts for fine-grained creative direction.
Edit lyrics and vocal melodies directly in generated audio. Symbolic representations let creators read and modify musical content.
44.1kHz stereo output with in-house Music Information Retrieval models for beat tracking, key detection, and chord analysis.
Seedmusic
on SeaDance AI
Access Seedmusic inside your creative workflow to generate professional music with ByteDance Seed-Music's unified generation technology.
Start CreatingSeedmusic is built for musicians, songwriters, producers, and content creators who need controllable, high-quality music from ByteDance Seed-Music technology.
Try Seedmusic
Convert scores to songs, iterate on lyrics and melodies, and explore new musical directions with controllable generation.

Leverage vocal music generation, singing voice synthesis, and post-production editing in one unified workflow.

Generate custom background music and vocal tracks with style control for social and brand content.
Experiment with Seedmusic's workflow and combine style descriptions with score references to create studio-ready music tracks.
Start Creating
Discover Seedmusic's advanced features powered by ByteDance Seed-Music's unified generation framework.
Start CreatingBuilt on audio tokens, symbolic tokens, and vocoder latents—each chosen to best support musical workflows from ideation to delivery.
Generate high-quality vocal music and instrumental tracks with performance controls from multi-modal conditioning signals.
Edit lyrics and vocal melodies directly in generated audio using symbolic representations creators can read and modify.
Low-threshold voice cloning for singing voice synthesis and conversion with professional 44.1kHz stereo output.
ByteDance's Music Information Retrieval models power beat tracking, key and chord detection, and symbolic feature extraction.
From ByteDance Doubao Seed team's Seed-Music launch to score-to-song workflows, Seedmusic empowers musicians, songwriters, and content creators.
Convert scores to full songs, edit lyrics and melodies, and iterate on musical ideas with controllable generation.
Generate custom background music and vocal tracks for social content with style-controlled output.
Leverage Seedmusic's unified framework for vocal music, instrumental tracks, and post-production editing.
Create on-brand music with audio references and style descriptions for campaigns and product launches.
Join creators using ByteDance Seed-Music technology to produce high-quality tracks. Seedmusic makes professional music generation accessible to everyone.
Powered by ByteDance Seed-Music