SongGeneration Studio is a slick interface for Tencent’s LeVo music model
SongGeneration Studio just landed on GitHub — an open project built around a friendly interface for LeVo, the music generation model from Tencent AI Lab that actually does the heavy lifting. The prompt lets you set lyrics, genre, mood, voice tone, instrumentation, and tempo, and you can lay out the track structure in blocks: intro, verse, chorus, bridge, outro, instrumental.
What’s under the hood
The tool can output a full mix or split it into vocals-only and instrumental-only stems, supports queued batch generation, and stores everything in its own library with export to FLAC or MP4.
A single track takes 3 to 6 minutes to generate — not instant, because everything runs locally on your own hardware instead of in the cloud.
Is it really a Suno replacement
That is really the catch: this is not a cloud service, it’s a local tool that needs a GPU with at least 10GB of VRAM and 25GB of free disk space — roughly 15GB of which goes to the model files during setup.
For anyone with the hardware to run it, that setup buys a lot more direct control over the track than cloud tools like Suno typically offer. Calling it a free replacement for Suno for everyone, though, is a stretch — the hardware bar is too high for a casual user to just try it on a whim.
What caught my attention wasn’t the generation itself, but the approach — instead of one text prompt for “mood,” you can dial in voice tone, tempo, and structure separately. But this is a tool for people with a serious GPU, not a “make me a hit” button for everyone.
Share
SUBSCRIBE TO OUR PRIVATE CASES AND USEFUL TIPS
Subscribe to our newsletter, get only exclusive content and weekly digests, no any spam!
By providing my email, I accept the Privacy Policy.