🌍 What if the entire Earth could be streamed as a photorealistic 3D world?
This prototype renders billions of 3D Gaussian splats directly in the browser.
Powered by @playcanvas . WebGPU + WebGL.
The technology is ready. The remaining challenge is capturing the planet.
👇
Here is a 3DGS scan of an aircraft and its engine, integrated into a @playcanvas experience. Users can freely explore the aircraft, inspect the engine, interact, and access photos and additional information.
After switching to @PlayCanvas' high-performace WebGPU renderer yesterday, some huge splats are starting to arrive on SuperSplat. Here it takes me a minute to speedrun the full length of this Swiss village. It's 730MB of SOG data yet it loads almost instantly and runs buttery smooth on my laptop. 🏃♂️⚡️
It's all free & open source. 💚
📖 Read the full breakdown: https://t.co/CjxGgRgC6p
💬 Join our Discord — where the world's best splat creators hang out: https://t.co/Dgjrkl1LGG
🚀 Major upgrades just landed in SuperSplat, the free and open source platform for 3D Gaussian splatting.
Here is a 24-MILLION-Gaussian scan streaming live to a web browser. Near instant load time. Solid 60 fps.
How? A new compute-based WebGPU renderer + automatic LOD streaming. 🧵
⚡ How fast is the new compute-based WebGPU renderer vs WebGL 2?
📈 Desktop: up to 5.7× faster on big scenes
📱 Mobile: ~2× faster
WebGPU now reaches ~85% of users; everyone else falls back to WebGL 2, rendering identically. Ships in PlayCanvas Engine 2.19.0.
Open source library splat-transform by @playcanvas just hit 10K weekly downloads on NPM. Check it out for all your 3D Gaussian splat processing needs! 🛠️
Over 10k views on my Gaussian Splat of the Moon! Next up, I hope to create a complete 3D representation of our celestial neighbor (and I’m just waiting for more high-res pictures of Mars 😅).
View it in 3D: https://t.co/5EMp9AlmKQ
@playcanvas@lichtfeldstudio@NASA
Enter the Backrooms - in 15 minutes
🪄 Upload a movie still to @SpAItial_AI Echo-2
⏬ Download a 3D Gaussian splat PLY file
🥽 Upload to @PlayCanvas SuperSplat and enter VR
[1 / 3]
Remember the famous locomotive splat from 2023? At the time, it was considered impressive. But look how @playcanvas handles a 10M Gaussian scene in 2026:
⏬ Streamed level of detail
🧊 Accurate voxel collision
⚡️ Compute-based WebGPU renderer
⏲️ Near-instant load times
[1 / 4]
Opus 4.8 is insanely good at vibe coding games using @playcanvas. This was made conversationally over an hour. I didn't look at the code once. Every request carried out perfectly. Game size: 280KB. Playable 🔗 in 🧵
I just one-shotted a 3D world with proper collision physics using the @SpAItial_AI API and @playcanvas
Here's how i did it:
-> Install the spaitial-playcanvas-world skill: npx skills add spaitial-dev/spaitial-playcanvas-world
-> Open Cursor or your preferred coding agent
-> Paste the following prompt or use your own:
"Using /spaitial-playcanvas-world skill and the spaitial api key provided build a PlayCanvas game from this world prompt:
An empty monumental desert sci-fi palace interior with colossal stone arches, sand-dusted floors, bronze industrial machinery, filtered sunlight through high slit windows, carved geometric walls, long ceremonial corridors, immense scale, cinematic warm shadows, ancient-futurist architecture, quiet unoccupied environment. No people, no humans, no characters."
This skill covers World and Collision mesh generation using the SpAItial api, and viewer + physics using Playcanvas.
Splats are getting larger and more detailed. The dataset for this scan of Koriyama Castle in Japan is half a gigabyte. But it loads nearly instantly. Imagine being able to visit perfect digital copies of any location in the world - this is where SuperSplat is heading.
[1 / 2]
Tourism, location scouting, cultural heritage, real estate...3D Gaussian splatting will disrupt them all.
This splat of Ferstel Passage in Vienna was captured on an @XGRIDS_OFFICIAL PortalCam and uploaded to @PlayCanvas SuperSplat. The detail is mindblowing. 🤯
[1 / 2]
splat-transform's offline rasteriser now supports depth of field.
Each Gaussian dilates by its own circle of confusion in the projection pass.
New flags: --f-stop, --focus-distance, --sensor-size.
This test was rendered with a simulated f/0.8 aperture: