Transforming 'Little Bo Peep' with Lullabize AI: Cognitive Training Music | The Lyrical Literacy

This demo pastes Little Bo Peep into Lullabize, restructures it for cognitive development, then reshapes the result through Suno with different styles and a cloned voice.

10:35 video4 min readWatch on YouTube

Turning a familiar nursery rhyme into a genuine cognitive-development tool is a different problem than just setting it to music. This walkthrough demonstrates Lullabize, a tool that restructures songs according to neuroscience and learning theory, using the classic "Little Bo Peep" rhyme as the test case, then pushes the result through several rounds of iteration to show just how much a generated song can change depending on the choices made after the first draft.

What Lullabize actually does

Lullabize is built to create simple songs and lullabies for cognitive learning, grounded in neuroscience and learning theory rather than just rhyme and melody for their own sake. The process starts by pasting in existing text, in this case the "Little Bo Peep" nursery rhyme, and giving the tool a command to "lullabize" it. What comes back is a song restructured for cognitive development based on that underlying theory, which the video frames as a first draft rather than a finished product. Alongside the restructured lyrics, the tool also generates suggested prompts for a companion video, useful if a creator wants to take the output further into a full production.

Treating the first draft as a starting point

A key idea running through the demo is that the Lullabize output is meant to be edited and reshaped, not used exactly as generated. The lyrics give a creator a structurally sound starting point, already built around cognitive-development principles, but the real personalization work happens afterward, when that draft gets fed into a music generation tool and pushed through different creative choices.

Iterating through Suno

The next stage takes the Lullabize output into Suno to see how much the platform can change the final sound. The first pass is a straight cut-and-paste with no modification to the Lullabize output, producing a pleasant but fairly plain, generic-sounding result. The second pass experiments with moving away from traditional lullaby sound entirely, mixing in styles like rap and country to change the character of the song. The video is candid that musical taste here is subjective: what sounds bland to one listener might be exactly what another prefers, and that's fine, since the tool is built to support many different outcomes from the same starting lyrics.

Adding a cloned voice for personalization

The third and most distinctive iteration involves uploading a recording of one's own voice into Suno before generating the song again. The claim made here is that doing this adds a kind of grittiness or authenticity to the output, in the same way that starting from a real photo and having it stylized still preserves something recognizably you, even after the image gets cleaned up. Uploading a real voice recording changes the generated voice while retaining some of what made the original recording distinctive, and this version is the one ultimately preferred over the two earlier attempts.

Personas, voice uploads, and a technical distinction

The video also draws a clear technical line between a Suno "persona" and what happens when you upload your own voice recording. A persona, in Suno's terms, can only be created starting from a song that wasn't itself generated from an uploaded audio file. Because the voice-driven version here started from an uploaded recording, it technically can't be turned into a persona, even though functionally it behaves the same way: a consistent voice profile used to generate further songs. The suggested alternative for anyone who doesn't want to upload their own voice is to generate a large batch of songs, pick a voice you genuinely like among them, and turn that one into a proper persona instead.

Key takeaways

  • Lullabize restructures existing text, including nursery rhymes, into songs designed around neuroscience and learning theory for cognitive development.
  • The tool's output is meant to function as a first draft, both for lyrics and for the video prompt suggestions it generates alongside them.
  • Feeding the same Lullabize lyrics through Suno with different style tags produces meaningfully different results, from a plain default sound to genre blends like rap and country.
  • Uploading a personal voice recording to Suno before generating a song changes the output's voice while carrying over some character from the original recording.
  • Suno's persona feature can't be created directly from a song generated using an uploaded voice recording, even though the practical effect is similar.

Who this is for

This demo is part of The Lyrical Literacy project from Humanitarians AI, which uses simple songs to strengthen brain areas tied to logical thinking, for children and adults alike. It's useful for anyone curious about AI music generation workflows generally, and especially for educators or parents interested in trying Lullabize themselves to turn a familiar rhyme or story into a personalized cognitive-training song. A related, more advanced tool called Musinique, aimed at guiding song creation through explicit learning rules, was in development at the time of this video.

Full transcript(auto-generated, with timestamps)

[0:01]Hey, it's Bear here. Uh, talking about the Lullabable software a bit more. Uh, again, Lullabableize. You can go to humanitarians.ai, go down to Lullabableize, and it'll get you a link to the lullabies tool. So, what the lullaby tool is, Lullabi tool is, is to uh create little simple songs for cognitive learning. you know lullabibis and symbol songs that are based on neuroscience and learning theory. So what you do is you I paste in little bow peep here gave it the command lise and then it generated a song based on sort of how it should be structured for cognitive develop based on theory and that's what this tool has by the way I'm

[0:51]Writing a different tool called musake for any kind of learning this is a little bit more complicated so this might take a couple weeks to finish um but I will also have the same idea. It will have rules about how you learn things through songs and help you create a song with those rules helping you guide the lyrics. After you do lyrics though, you can then start editing and changing. You should think of lyrics as sort of a first draft where you want to get started. Good thing about this first draft, it also gives you, by the way, suggestions for prompts for a video if

[1:26]You want to do that. You want to create a video. We'll do other things on videos. So what I'm going to do is now take this into suna and show you how much suno can affect the final output. So the first one I did here was just cut and paste. Did not do anything to the low level ice output. Stop. So, they're nice, but I find that kind of a little bland. But again, music is very, very personal. If you love that, great. Uh, what I then did is I just mixed up sort of not using traditional lii

[2:37]Sounds, but things like rap and country and all that in there to just change it up a bit, see what it sounds like. When you use AI, it's a very iterative playing. You play with things, try, play with things, try, play with things, right? It took me 10 minutes to create, you know, 20 songs. So, it's easy to just try things. Try things. Try things. Try things. So, let's let's see what this sounds like. littlest sheep. >> So, I like that a little bit better. I'm not in love with that, but I like it a little bit better than the traditional lullabi. Again, music is

[3:35]Very personal. Maybe you love the first one. Maybe you like or dislike this one. Third thing I did is what I usually do and that's upload various recordings of my voice. And my belief, and you may disagree with me hearing these things, is it adds a sort of grittiness to the generated music. Just like you would if you, you know, upload a picture of yourself and had it transformed to like a, you know, official office picture and just clean it up. It would still be you, but it would just be cleaned up and prettied up. And sort of that's what you can do with the music as well is if you

[4:15]Upload you, it changes you. It gives you a different voice, but it keeps some of what it uploaded in what it uploads. And so this is the song again with a little bit of changing of mixing other things in it. The difference here is I uploaded my sort of gritty nasty voice to it and then had it make it pretty because that's what it does. >> So that's the one I'm going to go with. I like that one much much better than the other two. But this is just a sort of discussion of if you just cut and paste this in uh you'll get something that sounds nice

[5:13]But pretty sort of genericy AIE. You can mix that up by just making it unusual just by the tags and things you do. And you can make it more unusual by literally just uploading your voice or your kid's voice to syno again when you create something there's a upload. So if I wanted to create I can just upload an audio here. This thing called a persona as well. Technically, what I do is not a persona because they don't allow me to upload my own voice and make it a persona. But in effect, it is a persona because I'm basically just making a song for my voice and using the

[6:02]Song that I create for my voice to make other songs. Uh, but technically it's not a persona. Persona you can only create like I could create a persona from this I believe like create persona because it didn't start from an uploaded audio. If it start from an upload all I did have to make persona. I'm not going to make a persona because I don't like this voice. If I love this voice I would make a persona of it. Um I mean that's one way of doing it. It's just creating a bunch of voices, finding one that you like and making personas. I would recommend that if

[6:38]You're not going to upload your voice, just make a shitload of songs and pick the ones that you love and create those into personas. Technically, when I upload my voice, it's not a persona. Uh although in effect, it's sort of the same basic idea. It just technically it's not because Sono doesn't allow me to do that. But they do allow me. I for example, I could make another song from this song. Now that this song is based on my voice, create another song and maybe that changes my voice a bit and the rendering of the other song made softer harder. Then what I can do is I can just create

[7:12]Another cover from it. What I cannot do is I cannot create a persona from it because it originally came from me uploading my voice. But that's just a small technicality um just to let you know. Okay. Um, what I'm going to do for the rest of this video is just play the version that I like. Um, but this is just more insights on how to use this uh Liby tool. So like, subscribe, do all that stuff. I can't tell where to find them. Leave them alone and they'll come home and bring their tails behind them. Little B Pe fell fast asleep and dreamt she heard them bleeding.

[8:06]But when she awoke she found it a joke for still they all were fleeting. That up she took her little crook determined for to find them. She found them indeed, but it made her heart bleed. For they'd left all their tales behind them. It happened one day as B peeped it stray into a meadow hard by. There she spied their tails side by side, all hung on a tree to dry. She heaved a sigh and wiped her eye. And over the hills she raised and tried what she could as a shepherd is should that each tail should be properly placed. She gathered the tails each fluffy and

[8:57]Fine and thought these sheep they're out of line. With thread and needles she started to sew, stitching tails on quick row by row. But soon she saw to her surprise a tail had somehow stitched to her thighs. Oh dear, she cried, "This can't be right." With a tail on her leg, she was quite the sight. She stitched through morning, stitched through noon, stitched by the light of the high hung moon, till all were attached, tails snug and tight. But the sheep were gone, not in sight. Then down the meadow there came in a dash, galloping fast in a sheepish flash. Each sheep looking bare, each sheep

[9:44]Looking proud, leaving B peep laughing though crying out loud. The sheep wag their tails fluffy and grand, proud of their tails like a marching band. But soon they grew bored as sheep will do and wandered off without a clue. The sheep wagged their tails, fluffy and grand, proud of their tails like a marching band. But soon they grew bored as sheep will do and wandered off without a clue. Oh sheep, dear sheep, you'll drive me mad. You leave me tail tired, exhausted, and sad. And so the peep headed back, hoping they'd learn to stay on track.

More videos

Humanitarians AI Lyrical Literacy Project