Inspiration
Traditional VR video apps still treat immersive headsets like flat TV screens. You open a player… and watch a rectangle floating in space. There’s no depth, no context, no interaction, and no true feeling of presence.
We asked a simple question:
What if videos weren’t something you watch, but something you enter?
That question led to Enter XR a spatial, portal-based experience that transforms videos into immersive environments enriched with interactive facts, AI-generated highlights, and intuitive hand interactions.
What It Does
Enter XR is a next-generation immersive entertainment experience that turns ordinary videos into explorable 3D micro-worlds.
Key Features
Portal Entry: A glowing mixed-reality portal appears in your room. Walk through it to “enter” the video world. 3D Curved Video Surface: The video wraps around you on a cinematic curved plane for depth and immersion. AI Fact Cards: When interesting moments occur, floating fact cards appear in spatial space beside the video (like the overlays in high-production podcasts). Hotspots & Highlights: Automatically triggered callouts show key moments or details within the scene. Hand-Tracked Controls: Pinch to scrub, wrist menu to toggle modes, hand ray to interact — no controllers required. Universal Video Support: Paste any public video link Youtube or any other video and Enter XR converts it into an immersive spatial experience instantly.
Enter XR is not a video player. It’s a video-world engine.
How We Built It
We built Enter XR using modern WebXR and Meta Horizon technology:
Frontend
- Meta Immersive Web SDK (IWSDK)
- Three.js + custom shaders for curved video surfaces
- WebXR hand tracking API for pinch interactions and spatial UI
- HTMLVideoElement for decoding standard MP4/HLS links
- CSS3D + texture mapping for hybrid video rendering in VR
Key Systems
Portal System: A procedural ring mesh with glow shaders, animated using time-based displacement and reflection maps. Video Environment: A curved quad generated with parametric geometry + live video texture mapped dynamically. Fact Cards Engine: Simple JSON-driven timeline triggers spatial 3D cards appear at predetermined timestamps. Hotspot Manager: Places floating markers relative to the video plane and animates them into the user’s field of view.
Tools Used
- WebXR Emulator for fast testing without a physical headset
Challenges We Ran Into
1. Mapping 2D Video Into 3D Space
YouTube-like players do not allow direct texture access, so we pivoted to supporting public direct video links, using HTMLVideoElement as a dynamic texture.
2. Spatial UI Placement
Fact cards must remain readable, not jitter, and avoid blocking the video. Designing a stable anchor system that floats cards just outside the curved surface took iteration.
3. Hand Tracking Variability
Pinch detection differs across devices. We built:
- fallback pointer mode
- stable ray interaction layer
- distance-based pinch filtering
4. Portal Entry Transition
Creating a smooth animated experience going into and out of the portal required blending shader animation with camera movement without breaking immersion.
5. Performance (60 FPS requirement)
We optimized by:
- using lightweight geometry
- reducing overdraw
- dynamic texture resolution
- offloading UI to CSS3D where possible
Accomplishments That We’re Proud Of
- We built a complete portal-based immersive video system in under a week.
- Hand tracking interactions feel fluid and futuristic.
- Fact cards appear exactly like high-production podcast overlays but generated automatically.
- The entry/exit portal experience is magical and consistently gets a wow reaction.
- We created a new way of consuming videos not watching but stepping inside them.
What We Learned
- Spatial UI requires real design thinking flat UI patterns break immediately in XR.
- Immersion isn't about complexity, it’s about presence, readability, and simplicity.
- AI doesn’t need to be heavy — even lightweight summary models can enhance storytelling.
- Hand interactions require redundant pathways (pinch, ray and fallback pointer).
- Testing XR interactions without a headset relies heavily on WebXR emulator discipline.
What’s Next for Enter XR
1. AI Director Mode (Full Version)
Real-time scene analysis:
- object detection
- key moment predictions
- dynamic highlights
- conversational “explain this scene” mode
2. Multi-User Shared Portals
Watch videos inside the same portal with friends.
3. Immersive Depth Layers
Use depth maps to expand 2D video into a pseudo-3D volumetric environment.
4. Smart Camera & Mood Lighting
Adaptive lighting that matches video ambience.
5. Creator Tools
Upload a video - Enter auto-generates:
- fact cards
- hotspot timelines
- highlight reels
Built With
- three.js
- typescript
- vite
- webxr
Log in or sign up for Devpost to join the conversation.