Inspiration

Many people feel trapped by their own imperfections, yet those flaws are merely standards shaped by society’s gaze. Beneath our clumsiness, vulnerability, and shyness lie some of the most gentle and beautiful parts of being human. But we spend so much time hiding them that we forget—they can shine too.

What it does

It expands our creative process by combining human direction with AI-generated visuals and sound, enabling new ways of storytelling.

How we built it

In the early stage, we used tools like MidJourney to establish the film’s visual tone. During the image-to-video phase, we combined different AI tools to complement each other’s strengths. For audio, Suno generated custom music, while ElevenLabs and traditional sound libraries were used to build the full soundscape. The entire workflow wasn’t a one-way process, but a loop of “human–AI collaboration”: we gave AI instructions, AI returned new ideas, and its unpredictability continually enriched our expression.

Challenges we ran into

AI can’t always capture the exact shots we imagine, so sometimes we need to take advantage of its strengths in camera movement and design our shots around that. Some subtle or delicate moments aren’t suited for direct AI generation, and instead require us to rethink the scene from a visual-concept standpoint. In the end, it’s about finding a way to express ourselves within the “uncontrollable,” and making the most of the strengths that AI does offer.

Accomplishments that we're proud of

People were able to see what we wanted to express, and we achieved a result that we’re genuinely satisfied with.

What we learned

By embracing AI’s strengths and its inherent randomness, we’re able to create work with greater freedom.

What's next for A Little Full

To share our work with more people, learn from the feedback, and create something better.

Built With

  • 11labs
  • adobe-after-effects
  • adobe-premiere-pro
  • chatgpt
  • comfyui
  • dramina
  • dreamina
  • elevenlabs
  • ffmpeg
  • midjourney
  • nanobanana
  • photoshop
  • pika-labs
  • suno
  • veo3
  • vidu
Share this project:

Updates