Achieving chord recognition with the highest possible accuracy is our top priority since the start of the Chord ai project back in 2018. We are using in-house state-of-the-art deep learning models trained on thousands of songs that we have meticulously labelled over the years. Our deep neural network models are optimised to run locally on your device to make Chord ai as responsive as possible and also able to work offline flawlessly.
Today Chord ai is able to process any song on-device within seconds, achieving an unprecedented chord transcription accuracy, beyond trained human level. Chord ai can recognize most common chords such as major, minor, augmented, diminished, 7th, M7th, sus2 and sus4 chords. Pro users can also enjoy a whole world of beautiful chords, including half-diminished, dim7, M7b5, M7#5, 6, 69, 9, M9, 11, M11, 13, M13th, add9, add11, add#11, addb13, add13, 7#5, 7b5, 7#9, 7b9, 69, 11b5, 13b9, combinations of them such as 9sus4, min7add13, as well as inversion, a.k.a. slash chords such as C/E or Am/C.
Chords can be played in many ways and “voicing” refers to the specific arrangement of notes for each way of playing a chord. On a piano, it is the full list of notes together with their octave. On instruments such as guitar, ukulele, mandolin, banjo etc., it is the specific position used to play a given chord. In 2022, after years of research and experiments, we have managed to release voicing and position recognition, which means that Chord ai not only recognizes chord names, but can also show you voicings and positions which match best with the song. Combined with Chord ai’s massive library of chord positions, you will be able to play any song with beautiful complex chords like a professional.