Video wird geladen...
Video konnte nicht geladen werden
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.
321,454 Aufrufe • vor 16 Tagen •via X (Twitter)
46 Kommentare

1/n The core idea behind Echo-3 is that reconstruction is treated as a form of generation. Rather than only stitching together what the camera saw, Echo-3 combines the capture with what it has learned about how real spaces look. The result is a photorealistic, spatially consistent 3D world.

Echo-3 is state of the art for 3D reconstruction quality. It preserves more detail and produces cleaner, more consistent geometry across a variety of spaces.

Echo-3 outputs are 3D native and ready to use in downstream applications. Output can be exported as SPZ, SOG, meshes, normals, and depth. This can be used for visualization, simulation, VFX, robotics, and spatial applications. One capture gives you multiple production-ready representations.

We’re also releasing a new capture app for Echo-3. Capture a space with your phone, upload it, and get a 3D world. Available now on the App Store Enterprise customers can also access Echo-3 via our API. Reach out to [email protected]

Explore worlds captured on phones and reconstructed with Echo-3. From houses, cafes and museums to laboratories and industrial sites:

if event spaces had this I would save so much time

Now you can make one 😀

🔥🔥

Great team effort - awesome results :)

a part of my inner child self, who was hooked by Nvidia's instant-ngp, is amazed looking at this. wondering when you'll be able to run this on a GPU at home

Turning phone captures into realistic 3D worlds is impressive!

Thanks - check out the quality on the 3D scans in our gallery: But you can also create them yourself with our app!

Amazing quality!

Looks super cool! World models are crucial for machines.

@grok whats the price of this?

Impressive.

Looks neat, but windows and transparent areas need a bit of work. This might help to refer:

Looks incredible!

Many thanks !!!

Why is there no free trial, at least one generation?

Upload videos to the app when?

soon 👀

exceptionnel, avec genie 4 on mange bien

the fact that it can do outdoor spaces too is the part nobody's talking about. everyone's hung up on interiors but thats where this actually gets interesting

Echo-3 + 3D world models 🔥 This is where phone capture starts becoming much more than content. You are basically turning real spaces into usable digital environments.

how can i try it ? where it is available

I suppose It Needs a depth camera

Deleted. Need to sign up for $50 monthly (first month free) to generate a single splat and it fell apart in a single hotel room pretty quickly.

phone captures to metric scale 3d worlds spanning multi floor buildings is a serious jump, the metric scale part is what makes it actually usable for real applications not just pretty renders

把重建和生成结合起来很有意思 🧩 导出 mesh / depth 时,能否同时标出哪些区域有拍摄证据、哪些主要靠模型补全?做可视化可以接受“合理补全”,用于机器人仿真时,这个区分会很重要。

Try it out. We have a mesh api coming soon!

谢谢补充!期待 mesh API 🧩 如果示例能展示遮挡区域的重建结果,并区分实测与补全,就更容易判断哪些场景适合接入仿真了。

Is it better than Genie 4?

How much is it? I don't see any pricing info.

pricing is available on the app. For API, please reach out to [email protected]

API?

yes, please send an email to [email protected]

Not available in Brazil :/

A friendly suggestion: you should take care that your name is always interpreted “Spaitial” insteal of “SPALTIAL”. Perhaps by using lower case more. I had trouble finding you on the app store for that reason.

Any plans on adding webUI support for non-iOS users?

Turning everyday phone footage into 3D worlds is impressive!

Very cool. Can developers use this in their own apps using the camera for scanning?

Open Source??

Can I send in 100m scale captures?

the ipad app is so sloppy

Metric-scale phone scanning has an ETH backstory: in 2013 Marc Pollefeys' group showed an app that builds a 3D model live as you move the phone, computed on the device itself. The inertial sensors supplied the absolute scale. #ComputerVision
