My daughter Jenna turns 18 next month. I wanted to do something for her birthday that wasn't just a card and a gift card, and since she's grown up with my dad jokes she knows I'll happily embarrass myself for a laugh.

So I wrote her a rap. Then I used one AI tool to turn it into an actual song, and another one to make a video of me lip syncing to it. Here's what I ended up with:

The rest of this post is how it came together.

Writing the lyrics

I wanted the lyrics to hit on the stuff that actually makes Jenna Jenna. She's a second degree black belt in Tae Kwon Do, she just won a state bowling championship, she passed her driver's test, and she's heading to F.I.T. in the fall to study fashion. There's also a running joke in our house about her mom not being thrilled about her driving on highways yet.

My one rule going in was to only use free tools that were simple to use. No subscriptions, no tutorials, no credit cards.

Here are the final condensed lyrics I came up with:

Yo, gather 'round, this news can't wait —
Our girl Jenna just hit the big one-eight!

Black belt in TKD — second degree, no cap,
One roundhouse kick — Cameron never saw that!

Two-handed style, watch her bowl —
State champ crown and total control.
Closed the finale, left the whole place shook —
"STEE-RIKE!" — straight outta Al Bundy's book.

Passed the driving test (finally — phew!),
That '09 red Jetta's rollin' with you.
Cruisin' Ridgewood like it's her parade —
Stay off the highways, your mom is afraid!

From Mom, from Jack, from Dad — all three,
So proud of the woman you grew up to be.
F.I.T. next fall — go live that dream,
'Bout to make the fashion world scream.

Eighteen years down — we've loved every one.
Now go strike 'em down. Your story's just begun.

If you've never seen the Al Bundy reference, here it is:

Turning it into a song with Suno

I googled for a free app that could turn lyrics into a song and Suno was the first result. You paste in your lyrics, describe the style you want, hit go, and about 30 seconds later you have a full song with vocals and instruments. Wild.

For the style prompt I went with:

I want to create a high energy up tempo rap song in the style of Weird Al Yankovic

I didn't want a serious hip hop track - I wanted a goofy dad-rap. I think the song I chose has the Weird Al style with the pace of a Pitbull song.

Suno gives you two variations per generation. My first pass was okay but too polished. I regenerated a handful of times before I got something that felt right.

My Suno workspace with the different takes I generated before picking one.

The original track was about 1:45 and it was pretty funny the whole way through, but that felt long for something I'd be sharing around as a short video. I trimmed it down to under a minute.

Attempt one: me lip syncing it myself

I'm a big fan of the TV show Lip Sync Battle with LL Cool J, and I honestly thought I could pull this off myself. How hard could it be?

I played the song in the background and recorded a video of myself with QuickTime on my Mac. Watching it back, there were way too many silly gestures and you could hear me singing out of tune to parts of the song. I couldn't help myself.

I showed it to my wife. Her exact review was "super corny and you'd better not let her friends see it!" She was right. Time to hand this off to a computer.

Finding a lipsync tool that actually worked

Going in, I figured a lip synced version of this song would look like something out of a JibJab video from 20 years ago - a static cartoon face where the jaw just opens and closes along with the audio. This is their classic "This Land" video:

I was not expecting much.

This part took over an hour of trial and error. I tried about 5 different services and every one of them required me to sign up with Google. Most were hit or miss - either the free tier was too restrictive to do anything real, or I didn't get enough credits to actually generate a full video.

I eventually landed on FreeLipSync. You give it a photo and an audio file, and it generates a video of the person in the photo lip syncing to the audio. Free, no signup, and it worked.

For the source photo, I started with this picture of me:

The original picture.

I dropped it into ChatGPT and asked it to make me look like a rapper. Here's the much younger looking version of me that it gave me back:

Rapper Jay, courtesy of ChatGPT. This is the image I actually fed into FreeLipSync.

Upload the picture, drop in the MP3 from Suno, wait about 5-10 minutes, done. I was impressed with how it makes my head move around, makes the words on the sweatshirt move and adds in some demonstrative hand gestures. Definitely not production quality, but definitely impressive for a free tier.

The catch with FreeLipSync was a 20 second cap per clip on the free tier. So I chopped the audio into chunks, ran each one through FreeLipSync separately, and stitched them back together. If you watch closely there are two spots where the transitions get choppy - right after the Al Bundy line and right before the final chorus. I didn't feel like fixing them.

Thoughts

Start to finish this was a few hours of work, and most of it was writing the lyrics and hunting down a lipsync tool that actually worked. The AI parts were the easy parts.

A few years ago I couldn't have made anything like this. I don't have any musical talent and I'm not about to hire someone to record a birthday rap. But an evening of messing around with free tools got me something my daughter is going to love (and be a little mortified by), and that's a pretty great trade.

Happy 18th, Jenna. We love you. Now go strike 'em down. Your story's just begun!