If you’ve ever wanted a personalized message from a favorite TV character (but couldn’t afford their Cameo fee), technology has the solution. A clever, if inconsistent, audio tool called Uberduck allows you to type any phrase and have a simulation of a famous voice “read” it out loud. Here’s how to use it.
This is an experimental web app, so you won’t find it on the App Store. Instead, visit uberduck.ai in Safari to view the tool.
The only real caveat is that in order to synthesize speech, you’ll need to log in using a Discord account. If you don’t have one, it’s free to register – but bear in mind this will give the maker of Uberduck access to your public Discord information. Log in and hit Authorize if you’re ok with that.
Once signed in, the interface is thankfully very straightforward. Using the drop-down boxes, first choose a category of person (e.g. Actors; Rappers; The Simpsons) followed by the specific voice you want to simulate (e.g. Patrick Stewart; Kanye West; Marge Simpson).
Most of the choices here are plucked from cartoons and video games, though you’ll find a handful of famous real-life voices in the mix too.
Next, simply type anything you want into the text box below and hit Synthesize. After a moment you’ll see playback controls appear further up the page. Hit play to listen to your synthesized recording. If you really like it, you can use the Download button to save a copy.
Note that the synthesized voice models are user submissions and the quality does vary from voice to voice. There are also limits to what fake speech trained using machine learning can do, so make sure to experiment. We found the web app absolutely nailed certain voices – but even when it was far from realistic, the results were still amusing.
So give it a go! The future is now. It’s just a bit janky.