Flshbck

Inspiration

We all have people we care about whose stories, memories, and perspectives deserve to be preserved after they pass away. Today, those memories are often scattered across photo libraries, videos, voice recordings, message threads, and cloud storage. We save them because they matter, but rarely have a meaningful way to revisit them.

Flashback was inspired by the idea of transforming those disconnected files into an experience people can step inside. Instead of simply flipping through a photo album, users can explore an immersive world that represents a loved one’s life, personality, relationships, and legacy.

What it does

Flashback is an AI-powered virtual reality storytelling experience that turns personal memories into interactive 3D worlds.

Users can upload photographs, videos, audio recordings, text-message exports, written stories, and spoken memories about a person. AI analyzes that material to identify meaningful places, events, interests, relationships, and themes.

The experience begins inside a virtual gallery. The user selects a photograph or meaningful object, which acts as a portal into a world inspired by that person. Inside the generated environment, the user can walk around, examine objects, hear narration, and activate different memory pathways that reveal stories or perspectives.

For our hackathon experience, users enter a gallery and select a photograph connected to Steve Jobs. The image transports them into a recreation of the Palo Alto garage associated with Apple’s early history, where interactive objects help tell the story of innovation, creativity, and humble beginnings. How we built it

We designed Flashback as a generative pipeline that converts uploaded memories into the different elements needed for an immersive experience.

First, the system accepts several types of input, including text, images, audio, and video. Generative AI processes this information and creates structured prompts for the environment, narrative, characters, objects, interactions, voices, and sound.

We used World Labs and Marble to generate the explorable environment, including a dimensionally informed recreation of the Palo Alto garage and its exterior. Tripo supports the creation of interactive 3D assets, while Splatverse and related spatial-generation tools help build realistic world geometry.

The VR application is developed through Android Studio and tested using the Pico emulator and headset platform. We use ElevenLabs for AI-generated narration and voice experiences, with generative audio tools such as Suno supporting atmospheric sound and background music.

Together, these tools allow us to transform archival media into a navigable, interactive story rather than a static collection of files.