Recently, there's been a stream of new AI text-to-audio applications similar to the surge of AI chatbots that emerged after ChatGPT was released. Keeping up with the trend, former Google Deepmind researchers have launched an AI music generator called Udio. 

Also: Generating music using AI in Copilot just got even better

On Wednesday, after months in closed beta, Udio moved to a public launch, enabling anyone to access the new music generator from the Udio website for free.

With Udio, you can create a song from a text prompt and customize all of the song's elements -- including its length, vocals, lyrics, and more. 

Despite the many AI music generators now available on the market -- such as Google's MusicFX, Stability AI's Stable Audio, and Suno AI -- Udio claims that its music generator can achieve better results. 

"There is nothing available that comes close to the ease of use, voice quality, and musicality of what we've achieved with Udio -- it's a real testament to the folks we have involved," said David Ding, co-founder and CEO of Udio.

Having tested many AI music generators, I put Udio to the test to see if these claims are justified. 

Getting started is easy: Sign in, type a description in the textbox of the song you'd like generated, choose if you want the song to be instrumental or have lyrics (which can be your own or AI auto-generated by the AI), and wait for the song to be generated. 

Also: What is Suno? The 'ChatGPT for music' generates songs in seconds

Inspired by my lunch, I entered the following prompt into Udio, "Create a song about the joy of lunchtime with a fun, EDM beat." 

The Udio website claims it takes fewer than 40 seconds to generate a song; however, in my experience, it took much longer, totaling eight minutes. When I contacted the Udio team about it, they said the longer than usual wait times were due to the "avalanche" of interest, but that the team is working to get it fixed shortly. 

Regarding the results, Udio output two versions of my prompt, each 33 seconds in length. The song snippets were impressive, sounding as though they were produced professionally and with all of the elements of a song present -- as you can hear below. 

I was given the option to extend the song further, adding an extension before or after the clip as a section, intro, or outro. I chose to add another section. However, after waiting for 20 minutes and nothing showing up, I gave up waiting. 

Also: Stability AI's new audio model creates even longer songs - here's how to try it for free

Considering the long waits are likely a server issue due to unexpected, launch-day high demand, and likely an outlier in the experience, I will leave out the model's speed in my initial evaluation. 

The quality of the generated music seems promising and sounded fuller and richer than other models I have used before. It was easy to get started and included more personalization options than others I have used. Once the model is up and running, we'll go hands-on again to see how it compares with other models, factoring in its speed. 

Artificial Intelligence

I asked Gemini and GPT-4 to explain deep learning AI, and Gemini won hands down

How to use ChatGPT's file analysis capability (and what it can do for you)

I tried Copilot Notebook: Microsoft's new AI tool offers two handy prompt features

What to know about Mistral AI: The company behind the latest GPT-4 rival

• I asked Gemini and GPT-4 to explain deep learning AI, and Gemini won hands down
• How to use ChatGPT's file analysis capability (and what it can do for you)
• I tried Copilot Notebook: Microsoft's new AI tool offers two handy prompt features
• What to know about Mistral AI: The company behind the latest GPT-4 rival