Mooncake Song

Inspiration

Growing up, the Mid-Autumn Festival was a moment of family reunion and sweet mooncakes shared under a full moon. As a bilingual parent, I wanted to recreate that sense of wonder for kids today—especially for children growing up far from their family’s cultural roots.

I realized that kids learn best through music, rhythm, and visuals, so I set out to turn a cultural tradition into a joyful kids’ music video that families everywhere could enjoy.

What it does

“Mooncake Song” is a fun bilingual kids’ song in Mandarin and English, introducing four classic mooncakes: Five-Nut Mooncake, Snow Skin Mooncake, Lotus Egg Yolk Mooncake, and Sesame Flaky Mooncake!

Through Look, Smell, Touch, Taste, and Hear, children explore the unique textures, aromas, and sounds of each mooncake.

The song includes interactive guessing parts, where kids sing and guess which mooncake is coming next, and ends with a warm message of family reunion for the Mid-Autumn Festival.

How we built it

• ChatGPT to draft the first draft of lyrics that I iterated upon more later to make it more catchy.
  
• Midjourney to create the image and video clips.
• Suno to create the music.
• CapCut to do the video production and music refinement.

Challenges we ran into

Mandarin pronunciation limitations: Suno struggled with accurate Mandarin tones. To maintain authenticity, we supplemented the vocals using CapCut’s AI voice, which offered more precise pronunciation.

Cultural image quality: AI-generated visuals didn’t always capture the subtle cultural details of mooncakes, festival attire, or traditional settings. This required additional curation, manual editing, and creative direction to get the visuals right.

Accomplishments that we're proud of

The Mooncake Song gained significant traction, attracting a high number of views, likes, and positive comments from families and teachers. It successfully blended education, culture, and fun, proving that AI-assisted creation can resonate across communities.

What we learned

AI is a powerful amplifier for creativity, especially when paired with human cultural knowledge and intention.

Tools like Suno, CapCut, and image generation models can get you 70–90% of the way there—but human judgment is essential for authenticity, nuance, and storytelling.

Kids’ media can serve as an engaging channel for cultural appreciation and heritage education.

What's next for Mooncake Song

Expanding the project into a series of festival-themed bilingual songs.

Built With

• capcut
• chatgpt
• midjourney
• suno