Excited to release our first public AI web app, powered by Apple’s open-source ML-SHARP model, capable of transforming a single image into a 3D Gaussian Splat with real depth understanding in seconds. Try it here → #AI #Apple #VR #GaussianSplatting

Wonderful, but pay attention to license of ML-Sharp... no commercial usage.

And that’s exactly why this is a free, non-commercial demo. As intended by the license if you read it carefully.

can't stop

I think you’re out of gpus :(

It's fixed now!

Even though you put a disclaimer this is still violating the commercial use license and model distribution clauses

For anyone who was getting a 429 error, it’s now fixed!

I'm getting a 429🥲 `modal-http: Webhook failed: workspace billing cycle spend limit reached`

We are currently phasing non-completing requests due for high traffic. We are working to solve the issue ASAP.

Shortest brief I’ve seen all week. Let’s get to it.Usually means the scope creep is going to be legendary.Gaussian splats are finally making 3D assets actually usable for those of us who never mastered Blender.

first try failed

This is so cool! Congrats!

Very cool. My second try got this result.

Hi, It seems to be failing during the 'Finalizing output' phase. It results in a timeout.

this looks seriously impressive. the potential for 3D modelling is huge, especially with AI making it so much faster. can’t wait to see what people create with it. innovation like this is what we need more of!

what's the point of this?

Future of VR experiences and spatial intelligence

Wow, that’s impressive! Curious how accurate is the depth mapping compared to traditional 3D reconstruction methods?

Top it up :)

bro, you need top up your account ----billing limit reached. unless the image i used was too big 848x840?

What kind of use cases are you imagining here? Film or game environments potentially?

i love this! just dm'ed you niccolo!

wow, this is epic and totally useful!

"Modal request failed: 429 - modal-http: Webhook failed: workspace billing cycle spend limit reached"

damn cool renderings :) !

You are using Apple's weights?

Absolutely Cool

Single image → usable 3D in seconds is wild. This is the kind of thing that quietly nukes whole pipelines.

This is cool! I think a tool like this where you could adjust camera/lens size/focal length would be invaluable for the right people

Great work!

Wowcongratulations

Dooooooooope!

Impressive innovation! Can't wait to try it out.

That’s super cool

very cool, can you upload multiple images to fill the scene with more data?

UAWWWWW, es increíble!!!, estoy jugando y me parece fantastico. tengo que descubrir como incorporarlo en Adobe After Effect. muy buena herramienta amigos, muy buena 👏👏👏👏

Amazing 🤩

Love this demo. I'm waiting for pairing gaussians with foundation VLM “world models” - one photo or text prompt could bloom into a metrically-true 3-D scene. Feels like we’re just one model upgrade away from “snap → explore”.