Polymath is a machine learning toolkit that converts any music library into a production-ready sample library. Point it at a folder of songs and it analyzes, slices and re-pitches the audio into organized, key-labeled samples that producers can drop straight into a track. Created by media artist and developer Samim, it applies ML audio analysis to a classic sampling workflow. It is an open-source project on GitHub with around 1,600 stars.
Traditional sampling means hours of manual chopping, pitch detection and file naming. Polymath automates the pipeline: it separates and segments audio into musical fragments, detects pitch and tempo, re-pitches fragments to a chosen key, and exports them in an organized folder structure with consistent naming. The result turns forgotten record collections and field recordings into fresh, playable instrument material. It runs locally in Python and builds on established audio ML libraries for analysis and processing.
Polymath is free and open source under the MIT license.
Music producers, beatmakers and sound designers who want to mine their own audio collections for fresh, ready-to-use samples automatically.
Polymath was created by Samim Winiger, a Berlin-based artist and researcher known for experiments at the intersection of machine learning and creativity. The tool reflects a sampling culture borrowed from hip-hop production: dig through old records, find the loops nobody noticed, and organize them so they are ready to play. Its output naming scheme keeps the original key and tempo in every filename, which keeps large generated libraries browsable inside any sampler or DAW.