Most writing about generative AI in music concerns composition — models that write songs, or stems, or arrangements. What actually ships in working software tends to be smaller and more practical, and VirtualDJ 2026 is a clean example.
Its new AI features include ShoutOut and Vibescripting.
ShoutOut
ShoutOut generates spoken shout-outs and announcements — the “drops,” the venue name, the birthday message, the this one’s for Marcus, the branded ident.
That’s a genuinely unglamorous and genuinely real job. Mobile and club DJs record this material constantly, and it’s tedious: every wedding needs the couple’s names, every venue wants its own tag, every event wants a fresh set. Recording it means a microphone, a quiet room, and doing it again when the names change.
Generating it means typing a name. For that specific task, this is straightforwardly labour-saving, and it’s easy to see why it shipped.
Vibescripting is the more interesting one
The name is doing a lot of work, and the concept behind it — describing an intent and having software generate the behaviour — is the more consequential feature.
Where ShoutOut generates an asset, Vibescripting generates the performance logic. That’s a different kind of delegation: not “make me a sound” but “decide what happens.” In a DJ context that shades toward the software making set decisions.
Whether that’s welcome depends entirely on what you think the job is. If DJing is selection and reading a room, software that makes those choices is competing with the practice rather than supporting it. If the job is running a six-hour function where nobody is listening closely for the first four, automation is a mercy.
VirtualDJ’s own demonstration of the ShoutOut feature.
The thing worth naming
A synthesised voice announcing a real venue to a real crowd is a small but genuine instance of a much larger question: the audience is being addressed by something that isn’t a person, without being told.
For a branded ident nobody cares, and nobody should. For “this one goes out to Sarah and Tom,” the sincerity is the entire content of the utterance, and a generated version of it is doing something different from a recorded one — even if it’s indistinguishable.
We’re not going to pretend that’s a crisis. It is worth noticing that the first place generative voice is reaching a live audience at scale is not art or film but the functional speech around a DJ set, and that nobody is likely to disclose it.
Why it’s a useful signal
Two reasons to pay attention even if you’ll never open VirtualDJ:
This is what adoption actually looks like. Not a model release, but a feature in a shipping product used by a very large number of working performers, aimed at the most tedious part of their job. That’s where this technology lands first.
The pattern will repeat. Every performance tool has an equivalent tedious asset-generation task — VJ idents, transition stingers, gallery audio-guide segments, installation voiceovers in six languages. Expect ShoutOut-shaped features in all of them, and expect Vibescripting-shaped ones to follow and be more contentious.