previral.
A NEW LENS ON VIDEO

There’s more
to every frame.

See the signals beneath the story. Explore how video, sound, and language relate to predicted brain activity — right on your Mac.

Build for macOS Take a closer look
BUILT FOR APPLE SILICON · macOS 26+
CORTICAL VIEWTRIBE v2
Cortical surface with illustrative orange and yellow activation regions
PREDICTED ACTIVITY20,484 VERTICES
+ INTERACTIVE ILLUSTRATION · SYNTHETIC SIGNALS
RESEARCH MEETS NATIVE PERFORMANCE
Meta TRIBE v2+Apple Core ML+Native SwiftUI

01 / A DIFFERENT PERSPECTIVE

From a moving picture
to a picture of response.

Bring your video. Previral connects its sights, sounds, and words to a synchronized map of predicted cortical activity.

01 — INPUT

Start with a video.

Open a local video file. Its frames, audio, and speech become three complementary views of the same story.

02 — ANALYSIS

Follow the signals.

TRIBE v2 models run through Core ML on your Mac, connecting multimodal features to predicted brain responses.

03 — EXPLORATION

Find a new perspective.

Scrub through your video alongside a cortical map and network timelines. See how the response changes over time.

02 / MULTIMODAL BY NATURE

One story.
Three signals.

A video is more than its visuals. Explore each input’s contribution, then see where their responses overlap.

SEPARATE INPUTS. SHARED TIMELINE.

What you see

Visual features, movement, and the changing scene.

VIDEO

What you hear

Sound, rhythm, and the texture of the audio track.

AUDIO

What’s being said

Spoken words and the language that gives them context.

TEXT

03 / MADE FOR THE CURIOUS

SOURCE AVAILABLE

Look a little deeper.

Build Previral from source and explore your own videos.
Your next perspective is already in the frame.

Get started on GitHub

YOUR MAC. YOUR VIDEO.

  • PlatformmacOS 26+ · Apple silicon
  • Build toolsXcode + XcodeGen
  • ModelsConvert locally with Python
Explore the conversion tools