Artificial intelligence can do more than just code or write, it can also create music. By combining lyrics with a chosen music style, AI-based tools are now able to generate complete songs that sound surprisingly professional. In this post, I will describe how I used Google Gemini and Mureka.ai to create an AI-generated track about a digital rivalry between Cisco and Huawei.
After generating the MP3 song, I convert it into an MP4 video so it can be uploaded to YouTube, since the platform doesn’t support MP3 files directly.
This guide shows how AI can turn simple text into music from writing lyrics to publishing a finished track online.
From Lyrics to Sound
My experiment started with Google Gemini, which I used to generate the song’s lyrics. I asked Gemini to create verses representing a rap-style “cyber showdown” — a lyrical duel between Cisco’s “ASA-Man” and Huawei’s “Cloud-Sec.” The result was The Cyber Showdown: Extended War, a multi-round exchange full of networking jargon and a touch of poetic bravado.
1. Using Gemini AI for Creating Lyrics and Music Style
Here’s the full text of the song The Cyber Showdown: Extended War
Round 1: Cisco's Opening (ASA-Man)
Yo, step aside, here comes the King of the Backbone,
I built the internet, you’re just a cheap clone.
I’m Cisco, born ready, from Route to the Cloud,
Your security's whispered, mine is shouted out loud.
I rock the ASA, the Firepower Threat Defense line,
We block malware zero-day, your code is just fine... for a wiretap.
You're the Great Firewall, tainted from the core,
They see a sweet deal, but we know what's in store.
Your gear's on a watchlist, the global alarm's ringing,
While my gear's got a CCIE's blessing it's bringing.
You chase the price tag, I chase the threat intelligence,
You’re talking 5G speed, I’m talking global resilience!
Round 2: Huawei's Response (Cloud-Sec)
A "cheap clone"? Please, you're the legacy, rusty and slow,
You call that boom-bap? That's the sound of your budget going low.
I'm Huawei, the future, we own the optical line,
Your old IOS is bloated, my eSight is sublime.
You're stuck in the past, with your billion-dollar acquisitions,
I'm running AI-powered analysis across all positions.
You talk threats, I talk scale—you can't deploy like us,
You charge triple the price for a fraction of the thrust.
"Tainted," you say? That's just fear from the West,
We innovated beyond you, and now you can't pass the test.
Your so-called security is just a complex GUI,
While I'm securing the IoT edge and the future of the free.
You're bulky and vintage, I'm lean, mean, and deployed,
Your certifications cost more than the features you've employed!
Round 3: Cisco Fires Back (The Code Trust)
Talos hunts harder than your entire R&D floor,
We patch the vulnerabilities that your team ignores!
You talk about eSight? That's just a slick presentation,
My code base is clean, audited by every nation... that trusts freedom.
You call my equipment heavy? It's enterprise-grade steel,
Built for decades of uptime, that's how true networks feel.
You talk about 5G, but your SDN solution is tied up,
In a web of political favors, your infrastructure is wired up.
The Silicon Valley backbone versus a state-mandated tool,
You're a supply chain risk, that's why your footprint feels so cruel.
You boast about price, but cheap gear comes with a cost,
It’s the data integrity that's ultimately lost!
Round 4: Huawei's Finisher (The Speed War)
"Integrity"? You should check your IOS versioning history,
Remember those backdoors? That’s your pathetic mystery!
The world's moving on, you're still relying on copper,
I'm pushing Cloud Fabric, my throughput's a showstopper.
We pioneered IPv6 adoption while you were still lagging,
Your security platform is just feature-staggering and dragging.
You think the market cares about your aging NAC framework?
I offer zero-trust native where your old systems won't embark.
Style Definition and AI Composition
Next, I needed a music style for the track. I again used Gemini to generate a short description of what kind of sound would best suit the lyrics. The AI suggested a “glitch-jazz spoken word” vibe. It is something between lo-fi hip-hop and experimental jazz.
Here’s the style definition I used:
Glitch-jazz spoken word: lo-fi hip-hop beats with dusty snares, Wi-Fi pings, and synth hums. Soft jazz chords float in D minor under a whispering voice. Tempo 80 BPM, reverb like inside a Pi case. Glitch pops, packet noise, and fading static—Wi-Fi gospel for the digital dreamer.
2. Using Mureka AI for Creating MP3 File
I took both the lyrics and style and pasted them into Mureka.ai, an AI-powered song generator. The tool synthesized two complete song versions as depicted in Figure 1:
Figure 1 - Mureka GUI - Two MP3 Songs Generated
Downloading an MP3 song for free from Meruka AI is a bit tricky, since this option is only available with the Basic Plan, which isn’t free. To work around it, right-click on the Play button (see Figure 2) and select Inspect.
Figure 2- Mureka AI song with Play Button
Next, open the Network tab, then select Media, and press Play. An MP3 file will appear in the list (Figure 3). Click on it, and the download will start automatically.
Figure 3 - Downloading MP3 file from Mureka AI using Inspect in Google Chrome
3. Uploading the MP3 Song to YouTube
YouTube doesn’t support direct uploads of MP3 files, so the track must first be converted into a video format (MP4). The easiest way to do this is by using tunestotube.com. This is a free online tool that automatically generates a video from your audio and uploads it to YouTube.
To use it, you need to sign in with your Google account, since the site connects directly to your YouTube channel. Once logged in, simply upload your MP3 file along with a background image for the video.
If you do not have a suitable image, you can quickly generate one for free using artist.io AI. Make sure the image resolution is 1920×1080 pixels, as that’s the standard YouTube video size.
After uploading both files to tunestotube.com, the site will automatically create the MP4 video and publish it on your YouTube channel — no video editing required.
Figure 4 - Final Youtube Video - Click Image to play music
Conclusion
Creating a song with AI was a great experience. Using Gemini for writing and style generation, and Meruka.ai for music synthesis - proved that you don’t need a studio, instruments, or a band to produce your own music.
However, there are also drawbacks. One ethical concern is that AI models may have been trained on copyrighted songs, raising questions about originality and fair use. Futhermore, while AI can reproduce musical styles, it still lacks the emotional intuition and subtle creativity of human composers.
Still, AI music tools like Meruka.ai open new ways for hobbyists, educators, and professionals to experiment with sound.