How to Create AI Music Video for FREE (Full Tutorial)

All right, so you want to create an AI

music video, but don't want to spend

hundreds of dollars on premium tools. I

totally get it. Most tutorials out there

assume you have unlimited budgets. Well,

for the last weeks, I've been

experimenting with free tools, and I

found a way to actually create an AI

music video while using completely free

tools.

[Music]

>> Now, to be honest, is it going to look

like a Hollywood production? No. But you

can absolutely create something

genuinely impressive that you'll

actually want to share on social media

without spending a single penny. So, I'm

going to show you the exact five-step

system I use to create AI music videos

for free. And this includes song

generation, image generation,

imagetovideo generation, video lip

syncing, and post-production. By the end

of this video, you'll have everything

you need to create your own

professionallook music video that people

will actually think you spent money on.

Now, you may already have your own song.

If that is the case, then you can skip

to the next step because for the people

that don't have a song yet, I'm going to

show you how to generate your own song

with AI. And for this, I'm going into

Sunno AI. Their free plan gives you 50

credits daily, which translates to about

downloading 10 songs per day. That's

actually insane value when you think

about it. So, after you've created your

account, you need to add in your prompt

to generate a song that matches your

vision. But this is where most people

mess up. You can't just type in

something generic like, "Make me a pop

song." No, you need to be specific. For

example, the prompt that I use looks

like this. Young woman singing about her

life, dreams, and love, soft rock, pop,

ambient, soulful, chill synth. That

specificity is going to give AAI more

guidance to work with. Now, it's going

to generate two versions for you. And

don't just pick the first one blindly.

Actually listen to both. I've had times

where the second option was way better

than the first. And if I hadn't checked,

I would have missed out on a much better

foundation for my video. So, I'm now

going to generate my song, and we can go

to our next step. I'm going to generate

my images that are going to be the

foundation of how the video will look

visually since after the image

generation I will use an imagetovideo

generator to turn the images alive. So

for creating the images I'm moving to

design AI and their free plan gives you

32 image generation credits daily that

reset every 24 hours. That might sound

like a lot but those credits disappear

faster than you think so you need to be

strategic. So the first thing we're

doing is creating our main character

portrait. This is crucial because

consistency is everything in a music

video. You can't have your singer

looking completely different in every

scene. That's just going to look

amateur. So, open Design AI and click

create a new project. By the way, all

the links of the tools I use are in the

description below if you want to follow

the process with me. Now, then go to the

text to image tab and select the design

realistic V3 model. I use the following

prompt. A photorealistic portrait of a

young American woman in her late 20s.

Pops soft rock singer from Wings of My

Tomorrow standing outdoors in natural

daylight. She wears relaxed, stylish,

modern clothing, perhaps a denim jacket

or casual chic outfit. Expression is

soft, reflective, and hopeful with a

light breeze moving her hair. Background

is softly blurred greenery or urban

backdrop with warm sunlight casting

gentle highlights. 8K ultra realistic

photography, cinematic soft lens,

natural colors, dreamy glow. Before you

click generate, make sure you set

generation mode to HQ and then create

the image. After that, I upscale it to

four times to get the highest quality

possible. And then I download it. Now,

yeah, it's going to use more credits

when using the HQ mode and to upscale

it. But this is your main character. So,

you want her to look good. Eventually,

doing this will save you so much time

and mental energy. Now take your song

and upload it to ChatgPT or Grock and

type analyze the song theme lyrics and

provide high quality prompts for

generating multiple B-roll scenes later

to be used in image to video generation

for generating B-rolls for the music

video. This will give you highquality

prompts that you can use in the

storyboard tab. I'm not even kidding.

This completely changed my workflow

because I can imagine when you look at

how detailed my prompts are that it is

quite overwhelming to figure this out

yourself. So, now that we have those,

switch to the instant storyboard tab and

set it to V1 mode. Then upload your

character portrait and generate scenes

with the specific B-roll prompts you

just created. Here are my examples. For

scene one, singer performing on stage

under soft spotlights, modern pop

inspired clothing, passionate

expression, cinematic lighting. For

scene two, woman walking alone on an

empty road, open fields, warm sunlight,

reflective mood. For scene three,

floating feathers or dandelion seeds in

golden light, dreamy macro shot,

ethereal atmosphere, and here are the

other five prompts that I generated on

screen. Now, for your opening scene,

return to the text to image tab and

generate your opening shot like this.

For example, a cinematic sunrise over a

calm city skyline, golden light,

breaking through soft clouds, dreamy

tone, lens flare, 8K photography,

shallow depth of field, emotional yet

hopeful atmosphere. After doing that,

download the result as a JPG. Make sure

to always include your aspect ratio in

the prompt itself. So add 16 by9

widescreen or whatever ratio you're

using. It helps the AI understand

exactly what format you want and

prevents those weird square or portrait

results that don't fit your video. So we

got our images and it's now time to turn

them into videos. So for that I'm moving

to Grog AI. Their Gro 3 plan gives you

20 video generations every 2 hours and

the limit on the duration of video

generation is 6 seconds fixed. Now, once

logged in to Groke AI, open the imagine

tab and upload your selected images and

set the video preset to normal. You can

add custom action or motion prompts to

create dynamic B-roll. Just as an

example, when I uploaded scene 8, the

ocean cliff scene, and use this prompt,

I got an amazing result. Standing

straight looking at the ocean with arms

wide open, camera angle zoom into the

ocean, then this gives you a great

result. But let me walk you through the

exact process for a few more scenes so

you can see how this works. For scene

one with the stage performance, I upload

that image and use a prompt like singer

moves slightly while performing. Soft

spotlight creates gentle shadows.

Microphone stays in position. For scene

three with the floating feathers, I

might use feathers drift slowly upward

in golden light. Gentle breeze effect,

dreamy floating motion. The key is being

specific about the type of movement you

want. If you just say make it move,

you're going to get random results that

might not match your vision. But if you

describe the exact motion, such as

camera slowly pushes in or hair moves

gently in breeze, you'll get much more

controlled and cinematic results. Simple

movements work way better than complex

ones. Don't try to make your character

do a backflip or some crazy

choreography. Think more about gentle

hair movement, slow camera pushes,

subtle gestures, maybe a slight turn of

the head. The AI is good, but it's not

magic. And when you ask for too much,

you get weird morphing and glitches

that'll ruin your shot. Another thing I

learned is to always preview your

uploaded image before generating.

Sometimes the image gets cropped weird

when you upload it. And if you don't

catch that, your video generation is

going to be based on a cropped version

that doesn't look right. Just take 2

seconds to make sure the image looks

correct in the preview window. Now, to

save you a lot of time, batch your video

generations. Don't just make one video

and wait 2 hours to make another one.

Plan out all your shots, generate as

many as you can in your first batch,

then go do something else for 2 hours,

and come back for the rest. But we got

our videos right now and it's now time

to lip-s sync a couple of them to our

song. So for this we're using

lip-sync.video and when you sign up you

get 500 credits daily. It takes only 9

credits to lip-s sync a video with voice

over with a total duration of 5 to 8

seconds. So you've got plenty to work

with. Now sign into lip sync.video and

open the lip sync tab. Upload video

scenes with clear facial visibility to

be lip synced like scene 1, scene 3, and

scene 5. They can work great for this

because only use video scenes where you

can clearly see the character's face.

Sounds obvious, right? But I wasted so

many credits trying to lip-s sync

profile shots or distance shots where

the mouth wasn't clearly visible. The

tool needs a clear front-facing view of

the mouth to work good. Make sure your

video file is in a supported format. MP4

works best in my experience. Then below

that, you'll see another upload area for

your audio file. This is where you

upload that trim section of your song.

The timing is crucial here. Your audio

clip should match the length of your

video clip as closely as possible. If

your video is 6 seconds long, your audio

should be about 6 seconds, too. Don't

upload a 30-second audio clip with a

6-second video. The tool gets confused

and the sync won't be accurate. Now,

upload a trim section of the music track

that's containing clear lyrics. Then,

generate the lip-s synced clips and

download them from the My Collection

tab. This tool works best with short

clips, like 5 to 15 seconds max.

Although it supports 10 minutes of lip-s

sync duration, it's really optimized for

shorter clips only. So, you're not going

to lip-s sync your entire song. Instead,

pick the key moments. The processing

usually takes a few minutes depending on

the length of your clip. Once it's done,

you can preview it right in the browser

before downloading. Always preview it

first. Sometimes the sync isn't perfect,

and you might want to try again with a

slightly different audio timing or a

clearer video shot. So, I've now got my

song, my B-roll videos, and the lip-s

synced videos, and it's now time to put

them together. I've done this in the

video editor in design AI which keeps it

easy to use a tool we already have but

you can use any free editor tool. I

easily added everything after each other

and this became my final music video.

[Music]

I walk the streets where shadows play.

Chasing whispers of yesterday.

I'm flying on the wings of my tomorrow.

[Music]

Leaving all my yesterday

behind.

Dreams unfold. I'll steal or borrow.

Love is the man.

[Music]

Like I said before, it's not the

greatest music video of all time, but

for doing this for free in a couple

hours compared to any real life or AI

paid process, this is some great value

to work with. The links to all these

tools are in the description below. And

if you do end up with some budget later

and want to see what a paid tool can do,

I've got a video about DesignAI's full

platform that'll blow your mind. The

most important thing is to just start.

Take these free tools, follow this

system, and make something nice.