AfterLight
Memory deserves a place to rest.
A reflective, memory-grounded experience. Share memories at your own pace. Ben, an AI reflection companion, helps you organize and find meaning in what you’ve shared.
Learn more →Umbrella Research builds ethical, human-centered AI — a reflective space for memories of a loved one, and living guides for the places people visit.
Umbrella Research is a small team focused on ethical, human-centered approaches to memory and AI.
Our technology powers two products: AfterLight, a reflective space for organizing and revisiting memories of a loved one, and ALIVE, an AI-powered avatar guide that brings physical venues to life through natural voice conversation.
We value transparency, user agency, and clear ethical boundaries in everything we build.
Memory deserves a place to rest.
A reflective, memory-grounded experience. Share memories at your own pace. Ben, an AI reflection companion, helps you organize and find meaning in what you’ve shared.
Learn more →Your venue, brought to life.
AI-powered avatar guides for physical venues. Visitors scan a QR code and talk to a lifelike digital guide — in their language, about the space they are in.
Learn more →We combine research-grade neural rendering, real-time audio AI, and privacy-first infrastructure into products that did not exist before.
3D Neural Rendering
Photorealistic 3D avatars from photographs using Gaussian Splatting — not pre-rendered video, not 2D filters. Real geometry, real-time rendering, running in a browser.
Gemini Live + LAM Pipeline
Bidirectional native audio conversation with real-time lip sync. Audio drives 52 blendshape channels through our LAM audio-to-expression pipeline — the avatar speaks with natural mouth movement, not canned animations.
EU-Only, Zero-Trust
All data processed and stored in the EU. Scale-to-zero GPU infrastructure means cost only during active use. No data leaves Europe, no biometric templates stored, full GDPR compliance by design.
Every design decision is guided by ethical principles. Our products are not simulations of consciousness — they are tools for reflection and guidance, constrained by design.
No avatar is ever created without explicit, cryptographically verified consent from the subject. Consent can be modified or revoked at any time.
Our deployment guidelines draw on input from clinical psychologists and human identity researchers to prevent dissociation, misrepresentation, or harmful use.
Rendered output is cryptographically watermarked, with a provenance API planned so any observer can verify whether content was AI-generated or human-captured.
We are establishing an independent ethics board — AI safety researchers, bioethicists, and civil rights advocates — to review high-stakes deployments.
Umbrella Research is scaling from incubation to market with two products, a first live venue deployment, and EU-native infrastructure built for GDPR from day one.
We’re raising seed funding to scale infrastructure, expand venue deployments, hire key research talent, and bring both products to market.
We are selectively onboarding research partners, early adopters, venue operators, and academic collaborators. Access to our platform and ALIVE venue deployments is available to qualified applicants.
Tell us about your work and how you’d like to collaborate. We review every application.