We’ve written about AI-curation startup Musiio several times this year: its CEO Hazel Savage pitched the company impressively at our NY:LON Connect conference in January as well as at Midemlab in June, raising $1m of seed funding in between for good measure.

Part of the company’s offering to customers (including labels and streaming services) is using AI to scan a catalogue of music and automatically tag tracks with useful metadata. It’s the kind of technology that can be hard to grasp without seeing it in action, so now Musiio has launched a feature to do just that, called ‘DIY AI’. It’s a demo of how the company’s tech works: people can paste in a YouTube URL for any music track, and see what tags the system creates – for example genre, BPM, key, mood etc. ‘Old Town Road’, for example, is 49% pop and 44% hip-hop – that’s the AI’s predicted probability about the tag; is in the key of A♭ minor; has a tempo of 135 BPM and ‘moderate’ energy; is roughly in the middle of the negative-to-positive spectrum; and is tagged with the moods ‘quirky’ and ‘powerful’. Billie Eilish’s ‘Bad Guy’, meanwhile, is 100% electronic and R&B; in the key of G minor; has a tempo of 136 BPM with moderate energy; and fits the moods ‘vibrant’ and ‘quirky’. It’s clearly a year of quirky pop hits! But Musiio will be hoping the demo attracts customers on the strength of its promise to do this kind of tagging at huge scale.