Inspiration

This project was inspired by the quiet moments of loneliness that often carry unseen beauty. I wanted to explore how a single emotion can evolve—how darkness can slowly open into warmth, hope, and light. The idea was to translate that emotional journey into a music video created entirely with AI.

What it does

Whispers into Light is an AI-generated music experience that blends voice, music, and cinematic visuals into a single emotional narrative. It tells the story of a heart moving from silence and isolation into clarity and illumination—using soft vocals, airy harmonies, and warm visual transitions through fields of light.

How we built it

The vocals and humming were generated using ElevenLabs, crafting gentle, emotional tones.

The visuals were created with OpenArt, generating cinematic shots of Saraswati walking, singing, and interacting with the wind and the landscape.

Scenes were arranged to follow a narrative arc: darkness → breath → movement → light.

Editing and final sequencing combined all elements into a smooth music-video flow.

Challenges we ran into

Capturing consistent character continuity across multiple AI shots.

Matching the emotional tone between audio and visuals.

Achieving natural movement—wind, hair, walking—within AI-generated frames.

Balancing stylization and realism to keep the narrative believable.

Accomplishments that we're proud of

Creating a fully AI-generated music video from scratch.

Achieving strong emotional coherence between the song and visuals.

Building a clear cinematic theme with consistent lighting, color, and character presence.

Showing how AI can support artistic storytelling in a gentle, human way.

What we learned

How to guide AI tools to maintain style, emotion, and continuity.

The importance of pacing and visual rhythm when telling an emotional story.

How lighting and wind simulation affect the mood in AI-generated scenes.

That AI can amplify creativity—but direction still matters deeply.

What's next for Whispers into Light

Expanding the project into a full-length visual track.

Exploring more expressive facial animations and singing gestures using AI.

Creating alternate versions for different moods (darker, warmer, orchestral, etc.).

Possibly extending the story into a short AI film featuring Saraswati.

Built With

  • adobe-premiere-pro
  • blender
  • capcut
  • elevenlabs
  • english
  • gpt-based-prompting
  • kling
  • openart
  • veo3
Share this project:

Updates