Algoriffix builds AI that understands music as structure, not just sound: rhythm, pitch, harmony, dynamics and the relationships between them. We don't want AI to make music for musicians. We want AI that understands music well enough to work with them.
Most AI music systems focus on generating. We wanted to build systems that listen and interpret.
A musical performance is more than an audio waveform. It contains notes, timing, harmony, dynamics and the individual characteristics of the instrument and performer. Algoriffix exists to make this musical structure accessible to machines in real time, so that technology can respond to what a musician is actually playing.
Our approach combines digital signal processing, machine learning and symbolic modelling. We build systems that work with explicit musical concepts, making their outputs interpretable rather than treating music as an undifferentiated stream of audio.
Our technology is psychoacoustics-aware and instrument-aware, recognising that different instruments and playing techniques produce and express sound in fundamentally different ways. Where appropriate, this extends to the physical and musical characteristics of an instrument — such as fretboard position and fingering on guitar and bass.
Our algorithms are developed in-house and designed to run locally on a CPU, with real-time, low-latency operation where the application requires it. The aim is simple: musical intelligence should be available directly where the music is being made, without sending the musician's performance to the cloud.
Transkr turns a live performance into MIDI in real time, and a recording into MIDI or MusicXML for further editing.
Pitch Perfekt applies the same listening principle to the voice — detecting pitch in real time and offering scale snapping and harmonisation, giving singers greater control over pitch while preserving the character of their performance.
These are early expressions of a broader goal: building artificial musical intelligence that can listen, understand and respond to musicians as part of the creative process.
Dr. Stanislaw Gorlow, Ph.D.
Founder of Algoriffix
© 2020–2026 Algoriffix AB