Ethan Ngam retweeted
Echo-3 is now available in the API.
Send a capture from a phone scan and get back a high-quality 3D world.
Export it as a Gaussian splat (SPZ, SOG, PLY), a mesh, or a floorplan.
Had the pleasure of beta testing Physicl Studio for @SpAItial_AI, really good quality data and helped tremendously for training our latest world model Echo-3! Huge congrats to the @PhysiclAI team! 🚀
Today we're launching Physicl Studio, and we're really proud of this one.
Tell it the situation your model gets wrong. You get it back as images or video, varied thousands of ways, with exact depth, normals, segmentation, boxes and camera pose on every frame.
5
258
Ethan Ngam retweeted
Today we're launching Physicl Studio, and we're really proud of this one.
Tell it the situation your model gets wrong. You get it back as images or video, varied thousands of ways, with exact depth, normals, segmentation, boxes and camera pose on every frame.
Ethan Ngam retweeted
iPhoneで撮影するだけで、空間を実寸スケールの3D化📱
🔹SpAItialが3D再構築モデル「Echo-3」を本日公開
🔹専用カメラや三脚なしで撮影し、実空間を再構築
🔹撮影できなかった部分もAIが補完。iPhoneアプリで利用可能
※寸法精度の独立検証は未確認
開発：SpAItial
情報元はリプ欄。
Ethan Ngam retweeted
Can a few minutes of iPhone video become a metric 3D copy of a real space? Echo-3 from @SpAItial_AI (@MattNiessner, @davnov134) launched today with an iPhone app.
It treats reconstruction as conditional generation: it keeps what the camera saw and generates the parts it missed so they match the surroundings. The outputs are 3D Gaussian splats, point clouds and meshes at metric scale, so distances match the real room. The phone's LiDAR and IMU only set the initial camera poses.
The pitch for robotics: physical AI is short of real-world 3D data, and if any phone can produce a metric digital twin, capture scales with phones instead of scanning rigs. @ethanngam from the team already scanned houses and simulated robots exploring them.
So far the comparisons are the company's own: sharper surfaces and no floaters compared with Polycam, KIRI Engine and Scaniverse. An enterprise API is offered too.
For contrast, Mundo from NTU (teapaw.org/_bbelousov/status/2113…) builds explorable, simulatable worlds from 1–12 photos and keeps physics explicit as code.
Ethan Ngam retweeted
The barrier to entry for 3D keeps plummeting. I’ve seen a lot of people pairing Astra with Blender for scans, but I think this will be stronger and much more photoreal.
Not to mention all the applications this opens up for Physical AI.
Happy to share that our new world model Echo-3 is out! Scan any place with your phone and our model generates a world for you to explore and share with others
Had a lot of fun scanning houses and simulating robots in them and watching them explore the place 😄
Ethan Ngam retweeted
Introducing Echo-3. A state-of-the-art world model.
Echo-3 turns phone captures into photorealistic, metric-scale 3D worlds. From a single room to multi-floor buildings, outdoor and public spaces.
Available today.
Ethan Ngam retweeted
We're hiring PhD Research Interns at SpAItial♥️
Work with us on cutting-edge generative AI, including image & video diffusion, VLAs/WAMs, world models, and 3D scene generators.
Publish a paper at a top-tier conference for your PhD.
Interested? Apply here👇
Ethan Ngam retweeted
We're hiring at SpAItial 🚀🚀🚀
Want to build AI that generates 3D worlds and understands their physics? 15 roles in London & Munich:
🧠 RS: World Models · 3D Diffusion · 3D Recon · Robot Learning
🛠️ RE: World Models · 3D Recon · Graphics
⚡ ML Infra · Inference
Join us 👇
Looks really good, this will make scanning houses way faster and cheaper for real-estate listings!
Large, consistent, and high-quality 3D worlds from simple iPhone recordings.
Works for multi-room scenes, fills missing gaps consistently while maintaining high quality, and it’s super easy to use.
Now in beta with partners and coming soon to the SpAItial API.
Link to the scenes in the comments so you can navigate this world yourself 👇🏽
Ethan Ngam retweeted
Large, consistent, and high-quality 3D worlds from simple iPhone recordings.
Works for multi-room scenes, fills missing gaps consistently while maintaining high quality, and it’s super easy to use.
Now in beta with partners and coming soon to the SpAItial API.
Link to the scenes in the comments so you can navigate this world yourself 👇🏽
Ethan Ngam retweeted
Real-world capture is almost never complete. A missed corner, object, a room that didn't get enough viewpoints etc. leaving empty holes where the geometry should be.
Our SOTA model Echo is built for that reality. It fills unscanned regions from surrounding visual context, so you still get a complete, explorable 3D world instead of broken voids.
Incomplete capture shouldn't mean an incomplete experience.
Ethan Ngam retweeted
Our SOTA world model, Echo, reconstructs high-quality scenes from sparse inputs and fills missing areas using its world knowledge.
That means far less capture, data, and prep to produce geometrically consistent, multi-room reconstructions.
We can’t wait to share it. Public release coming soon
Ethan Ngam retweeted
Announcing the SpAItial Startup and Creative Partners Program for teams and creators building with world models.
Accepted applicants get free credits, priority API access, early Echo model access, integration support, a direct feedback loop with the lab.
We want real builders using Echo in real workflows so your feedback shapes what comes next.
Application links in the comment 👇
Ethan Ngam retweeted
Sparse captures in, complete worlds out.
Echo reconstructs scenes from a small set of views, using world modeling priors to fill what sparse 3DGS leaves behind.
Coming soon to the SpAItial API.
Ethan Ngam retweeted
New from SpAItial. You can now:
- Create 3D world directly from the Claude desktop app using the new SpAItial AI plugin
- Create worlds with your agents using the SpAItial MCP. Works for Claude Code, Codex, OpenCode, VS Code, Windsurf etc
Create, edit, refine your worlds without leaving your chat! Links below 👇
Ethan Ngam retweeted
Fun weekend project: using @AIatMeta Sam3 to segment objects inside a @SpAItial_AI scene.
Pipeline:
-> Generate a world on @SpAItial_AI app and export as .ply
-> @playcanvas viewer to render splat + a uniform spatial grid over splat centers
-> Capture frames + camera pose, POST to a local SAM3 server
-> Lift each 2D mask → 3D: frustum-cull splats via the grid, reproject with the pose
-> Recolor / isolate the segmented Gaussians
This is all running locally on my Mac. Static scene, segment once, reuse persisted mask.
All open source (link in comment), you can clone and build on top of this 👇🏽
Ethan Ngam retweeted
SplatArena: Multiplayer arena is LIVE 🚀🚀🚀
Quake3-style combat. Drop in, frag your friends, and play on a world generated by Echo-2.
Play now, directly in your browser:
arena.spaitial.io/