I usually try to come up with one new crypto idea every couple of weeks and build it out. This time, I couldn't come up with anything. So instead of forcing it, I decided to do something different.

I made a music video.

Now, I’m not a video maker by any means. I’ve got basically zero experience with editing, design, or anything artsy. I’m more of a right-side-of-the-smooth-brain Solidity dev.

But that’s precisely why this project was so fun. It proved to me that anyone can build anything with AI.

In this article, I’ll walk you through how I made the video step by step, the tools I used, and how you can make your own, too.

The idea? A grunge cat band called Purrvana.
The lyrics? Only words cats can say.

The video introduces the band’s lead singer, an orange cat with a tragic backstory. He sings grunge because he lost the love of his life, his soulmate.

Before I get into how I made it, here’s the final video 👇

Everything starts with the song. It has to be a banger that makes you feel something; otherwise, the rest of the project becomes harder (cause you're gonna be listening to it on repeat as you build the video).

I used Suno to make the song. My first prompt was pretty basic: “A grunge rock song sung by a cat. Only meows.”

The early outputs were rough.

So I went to Gemini and asked it to create a better prompt for Suno with all the details I wanted. Tone, emotion, structure, instruments, vocals, and even the energy level. It came back with a long, detailed prompt.

Once I used that, Suno started making some good songs. The new songs actually sounded like something a cat grunge band would release. After a few attempts, I found the one. That became the first official Purrvana track.

With the song ready, I just listened to it on loop and tried to imagine what kind of story it was telling.

Here’s what I came up with.

It’s about an orange cat who used to be happy. He meets a girl cat, and they fall in love instantly. We see them having fun together, sharing snacks, getting into a little mischief, and even playing pranks on each other.

But the orange cat has a problem. He can’t stop pushing things off tables. Vases, cups, whatever’s near the edge, he wants to push them off.

One day, he pushes a lit candle off the table. It lands on some paper, and the whole house catches fire.

He runs out of the house, panicking, and finds his love outside. She has no idea what happened. But when the humans come out and see the mess, they realize it was them.

They chase the cats, trying to catch them. The two lovers get split up. She’s captured and taken away. He’s left alone, heartbroken.

The screen fades to black. Then the word Purrvana appears.

That’s the band’s origin story.

Once I had the story, I went back to Gemini. I told it to break the story into a list of scenes and attach a video prompt to each one.

Each scene needed to match the tone of the music and stick to a consistent art style. Gemini gave me a breakdown of every moment, with camera directions and moods, all ready to use as video prompts.

The hardest part of AI video making is keeping characters consistent between clips. Every video generation is isolated, which means one scene’s cat might look totally different in the next. The same thing applies to the art style of the video.

I used Veo 3, which has a feature called ingredients to video that lets you upload reference images and use them as “ingredients” to keep your characters consistent.

So I had Gemini write image prompts for the two cats: the orange boy cat and the white girl cat. Then I used Nano Banana to generate their character images. Once I had photos that I liked, I uploaded them as ingredients along with each video prompt.

That locked their appearances across scenes so the whole thing felt like one continuous story instead of a random collection of clips.

For each scene, I uploaded:
• The video prompt
• The boy cat image
• The girl cat image (if she were in that scene)

Then I let Veo 3 generate the clips.

Some came out great, others not so much. That’s just how it goes. Sometimes the art style was off, sometimes the scene didn't make sense.

When that happened, I’d tweak the prompt or re-run it a few times until I got something usable. Occasionally, I’d even rewrite a scene if AI couldn’t pull it off visually.

It’s all trial and error, and every time you do it, you get better at knowing what the AI likes. For example, you realize that you need to have the art style of the video input repeatedly in each video prompt; otherwise, they will all look different.

Another cool thing you can do is tell Gemini to add some cool camera angles/movement to the prompts. I know nothing about directing, so I just let Gemini have at it and do what made sense to it.

Once I had a folder full of clips and the final song, it was time to assemble everything.

I used Canva for editing. It’s simple, fast, and honestly great for beginners like me. I lined up clips to match the beat, timed emotional moments to the song, and added transitions to make the story flow.

That’s it. I now have a full music video for Purrvana

Here’s the complete tool stack I used:

• Suno for music
• Gemini for storyboarding and prompts
• Nano Banana for character design
• Veo 3 for video generation
• Canva for editing

This project was a good reminder that AI lets you do things you never thought you could. I’m a developer, not an artist, but I still made something that feels like art.

If you’ve been wanting to create something but don’t know where to start, this is your sign.

Pick a weird idea and build it.

Mess around and have fun.

You’ll figure it out as you go.

Thanks for reading!!
heeshillio

I usually try to come up with one new crypto idea every couple of weeks and build it out. This time, I couldn't come up with anything. So instead of forcing it, I decided to do something different.

I made a music video.

Now, I’m not a video maker by any means. I’ve got basically zero experience with editing, design, or anything artsy. I’m more of a right-side-of-the-smooth-brain Solidity dev.

But that’s precisely why this project was so fun. It proved to me that anyone can build anything with AI.

In this article, I’ll walk you through how I made the video step by step, the tools I used, and how you can make your own, too.

The idea? A grunge cat band called Purrvana.
The lyrics? Only words cats can say.

The video introduces the band’s lead singer, an orange cat with a tragic backstory. He sings grunge because he lost the love of his life, his soulmate.

Before I get into how I made it, here’s the final video 👇

Everything starts with the song. It has to be a banger that makes you feel something; otherwise, the rest of the project becomes harder (cause you're gonna be listening to it on repeat as you build the video).

I used Suno to make the song. My first prompt was pretty basic: “A grunge rock song sung by a cat. Only meows.”

The early outputs were rough.

So I went to Gemini and asked it to create a better prompt for Suno with all the details I wanted. Tone, emotion, structure, instruments, vocals, and even the energy level. It came back with a long, detailed prompt.

Once I used that, Suno started making some good songs. The new songs actually sounded like something a cat grunge band would release. After a few attempts, I found the one. That became the first official Purrvana track.

With the song ready, I just listened to it on loop and tried to imagine what kind of story it was telling.

Here’s what I came up with.

It’s about an orange cat who used to be happy. He meets a girl cat, and they fall in love instantly. We see them having fun together, sharing snacks, getting into a little mischief, and even playing pranks on each other.

But the orange cat has a problem. He can’t stop pushing things off tables. Vases, cups, whatever’s near the edge, he wants to push them off.

One day, he pushes a lit candle off the table. It lands on some paper, and the whole house catches fire.

He runs out of the house, panicking, and finds his love outside. She has no idea what happened. But when the humans come out and see the mess, they realize it was them.

They chase the cats, trying to catch them. The two lovers get split up. She’s captured and taken away. He’s left alone, heartbroken.

The screen fades to black. Then the word Purrvana appears.

That’s the band’s origin story.

Once I had the story, I went back to Gemini. I told it to break the story into a list of scenes and attach a video prompt to each one.

Each scene needed to match the tone of the music and stick to a consistent art style. Gemini gave me a breakdown of every moment, with camera directions and moods, all ready to use as video prompts.

The hardest part of AI video making is keeping characters consistent between clips. Every video generation is isolated, which means one scene’s cat might look totally different in the next. The same thing applies to the art style of the video.

I used Veo 3, which has a feature called ingredients to video that lets you upload reference images and use them as “ingredients” to keep your characters consistent.

So I had Gemini write image prompts for the two cats: the orange boy cat and the white girl cat. Then I used Nano Banana to generate their character images. Once I had photos that I liked, I uploaded them as ingredients along with each video prompt.

That locked their appearances across scenes so the whole thing felt like one continuous story instead of a random collection of clips.

For each scene, I uploaded:
• The video prompt
• The boy cat image
• The girl cat image (if she were in that scene)

Then I let Veo 3 generate the clips.

Some came out great, others not so much. That’s just how it goes. Sometimes the art style was off, sometimes the scene didn't make sense.

When that happened, I’d tweak the prompt or re-run it a few times until I got something usable. Occasionally, I’d even rewrite a scene if AI couldn’t pull it off visually.

It’s all trial and error, and every time you do it, you get better at knowing what the AI likes. For example, you realize that you need to have the art style of the video input repeatedly in each video prompt; otherwise, they will all look different.

Another cool thing you can do is tell Gemini to add some cool camera angles/movement to the prompts. I know nothing about directing, so I just let Gemini have at it and do what made sense to it.

Once I had a folder full of clips and the final song, it was time to assemble everything.

I used Canva for editing. It’s simple, fast, and honestly great for beginners like me. I lined up clips to match the beat, timed emotional moments to the song, and added transitions to make the story flow.

That’s it. I now have a full music video for Purrvana

Here’s the complete tool stack I used:

• Suno for music
• Gemini for storyboarding and prompts
• Nano Banana for character design
• Veo 3 for video generation
• Canva for editing

This project was a good reminder that AI lets you do things you never thought you could. I’m a developer, not an artist, but I still made something that feels like art.

If you’ve been wanting to create something but don’t know where to start, this is your sign.

Pick a weird idea and build it.

Mess around and have fun.

You’ll figure it out as you go.

Thanks for reading!!
heeshillio