Upload a song, pick your model, and our cluster generates a full beat map in minutes.
No cluster access needed — just drag, drop, and dance.
SCORE
0
COMBO
0
01 — Upload & Configure
🎵
Drag & Drop Your Song
or click to browse files
.MP3.WAV.FLAC.OGG.M4A
No file selected
Select Model
MuMu-LLaMA
Tested with Director/Actor approach. Acts as Director predicting rhythmic clusters,
translated into beatmaps by the Actor sub-decoder.
RECOMMENDEDHIERARCHICAL
Qwen2-Audio-7B
Multimodal onset extraction and hierarchical beatmap generation pipeline (similar to
MuMu) to compare rhythmic capabilities.
HIGH ALIGNMENT
Gemini (Pro / Flash)
Used for onset detection and benchmarking. API Key required! Please
change your API key after 30 mins to 2 hrs of use. Alternatively, install the project from our GitHub to run
securely in your own environment.
CLOUD API
DeepResonance & Flamingo
Multimodal models probed and measured for pattern finding and precision alignment
during benchmarking sweeps.
EXPERIMENTAL
⚡ Selected: DDPv2-Turbo
Song is sent securely to our cluster. You have no direct cluster access — results return here automatically.