Your products now have two shoppers: people and AI agents. Voxelo turns a three-minute phone video into a photoreal 3D twin in around an hour, and both can read it: interactive 3D, place-in-space AR with nothing to install, AI product and lifestyle imagery, and 2-3x the product attributes of a typical product page, published as JSON-LD. Already have splats? Host 100 free with viewer, AR links, editor and API. New in 3.0: API, analytics, publish to SuperSplat. Captured, not generated.
Hi Product Hunt 👋
I'm Ben, co-founder and COO of Voxelo. We're a small team in Manchester, and we've spent two years making one thing work properly: turning a short video of a real product into a photoreal 3D twin.
The idea behind 3.0 is simple. Your products now have two shoppers. A person, scrolling your product page. And an agent, reading it on their behalf. Both want to see the real product, and a flat JPEG serves neither of them well. A photo shows one angle. An agent can look at a photo, but it can't rely on one: no materials, no construction, nothing structured to match against the question it was asked.
So 3.0 opens two doors into one account, and both are free to start.
Already have splats? Bring them. Archives, heritage collections, product libraries, client work at volume. Host up to 100 of your own for free, made anywhere, with a viewer that does them justice, place-in-space AR links you can send to anyone, the editor, analytics and the API. Drag a PLY in; the Voxelo AR app opens SOG, SPZ, PLY and SPLAT natively. Publish to SuperSplat is live.
No 3D yet? Capture it. Film the product for about three minutes on a phone, rig or lightbox, and UG3D® (our own take on Gaussian splatting) turns it into a photoreal 3D twin in around an hour. 10 3D twins free. From that one twin: interactive 3D, hotspots, place-in-space AR with nothing for the shopper to install, AI product imagery and lifestyle imagery.
Then the part that's new in 3.0, and the reason we opened with agents: every twin publishes its product attributes as JSON-LD. Materials, construction, styling, details. 2-3x more than a typical product page carries, and all of it captured from the real thing rather than generated from a guess.
What it's not. Splats won't replace meshes. They capture and visualise; meshes design and edit. We're at our best on the materials that break 3D pipelines: suede, knitted mesh, stitching, weave, worn and weathered surfaces. The Virtual Studio is in beta and open to applications. Be curious about it, don't buy for it.
We'd value feedback over anything else today:
- Splat people: drag a PLY in and tell us where the viewer falls short.
- Ecommerce people: capture your hardest product and tell us what broke.
- Anyone building shopping agents: pull the JSON-LD and tell us what's missing.
You can try a live twin without signing up: https://app.voxelo.ai/viewer/md7.... Open it on your phone and place it in your room.
Vlad, Roman and I are in the comments all day. Ask us anything.
Ben
P.s. The boot in the film is a suede Chelsea boot. Suede is exactly the kind of surface a mesh pipeline hates. That's the point.
An hour from phone video to 3D is wild for small ecom teams. Does it fall apart on glossy or glass products? That's usually where photogrammetry gives up.
So good @ben_mckay , Vlad and team! Great to see Voxelo in action two weeks ago in Reading - so many use cases for ecommerce
Roman here, CTO, here to share a few technical details. 🤓
UG3D® (User Generated 3D) is built in part on 3D Gaussian splatting. Capture needs around three minutes of video, or a set of images, on a phone, a rig or a lightbox. Reconstruction takes around an hour, and every twin comes out of a pipeline tuned for the web, so a whole catalogue stays light on a phone.
Now, you can also bring your own! Pull in the splats you already have, from any pipeline. Each gets a viewer and a place-in-space AR link, and you can take it out again as PLY or SOG, or publish straight to SuperSplat. Your scenes stay yours.
Then there's attributes: derived from the product, adjustable by hand. Materials, construction, styling, details, made available via JSON-LD and the API, which is on the free account. This is the part that makes a product page legible to an agent, so it can find you e.g. a jacket that fits properly for your body type.
Splats capture and visualise, but they're specialised to different use cases vs meshes. But the adoption of splats for creative workflows seems to be on the up and up...
Upload the ugliest splat you have and tell me where the viewer falls over. That'll help me figure out what needs fixing next! 🚀
Love that it captures real textures instead of generating them. And making products easier for AI shopping agents to understand is a pretty interesting angle too.
@vladimir_mulhem This could be a game changer for e-commerce product pages. Can the generated 3D models be embedded directly into existing Shopify stores ?
Hey Product Hunt 👋 Vlad here, CEO and co-founder of Voxelo.
I built and sold a 3D agency, so I've spent twenty years wanting good 3D to be faster, cheaper and easier to share. Voxelo 3.0 is us opening the doors.
Our 3D twins are Gaussian splats, a new 3D format that captures the real product as the camera saw it instead of modelling it. Two ways in, one free account. Bring the splats you already have (we host 100 free, each with a viewer, a link and place-in-space AR), or capture a product for three minutes, on video or as a set of images, and get a photoreal UG3D® (User Generated 3D) twin in around an hour.
The new bit I care most about: one capture now does every output. Photoreal 3D, AR, AI imagery, and product attributes read off the scan and published as JSON-LD, which is what AI agents read. Your product page has two shoppers now, a person and the agent shopping for them, and one capture serves both.
Go bring a splat. Tell us what breaks. We're here all day. 🤘
Coming from the SEO side, the JSON-LD part caught my eye. How does it end up on the shop's product page? Do you hand over markup that goes into the HTML, or is it added with JavaScript?
Asking because we just looked at a few hundred AI-built sites, and on a lot of them the content only showed up after JavaScript ran. Most AI crawlers don't run JavaScript, so they never see it.
Putting materials and construction details into JSON-LD gives the 3D capture another useful job. An agent needs those details to explain why a product fits what someone asked for.
About Voxelo 3.0 on Product Hunt
“Phone video to photoreal 3D & JSON-LD in an hour”
Voxelo 3.0 launched on Product Hunt on October 22nd, 2026 and earned 93 upvotes and 13 comments, placing #13 on the daily leaderboard. Your products now have two shoppers: people and AI agents. Voxelo turns a three-minute phone video into a photoreal 3D twin in around an hour, and both can read it: interactive 3D, place-in-space AR with nothing to install, AI product and lifestyle imagery, and 2-3x the product attributes of a typical product page, published as JSON-LD. Already have splats? Host 100 free with viewer, AR links, editor and API. New in 3.0: API, analytics, publish to SuperSplat. Captured, not generated.
Voxelo 3.0 was featured in Artificial Intelligence (481.7k followers), Augmented Reality (44.8k followers) and E-Commerce (42k followers) on Product Hunt. Together, these topics include over 153.5k products, making this a competitive space to launch in.
Who hunted Voxelo 3.0?
Voxelo 3.0 was hunted by Ben McKay. A “hunter” on Product Hunt is the community member who submits a product to the platform — uploading the images, the link, and tagging the makers behind it. Hunters typically write the first comment explaining why a product is worth attention, and their followers are notified the moment they post. Around 79% of featured launches on Product Hunt are self-hunted by their makers, but a well-known hunter still acts as a signal of quality to the rest of the community. See the full all-time top hunters leaderboard to discover who is shaping the Product Hunt ecosystem.
Want to see how Voxelo 3.0 stacked up against nearby launches in real time? Check out the live launch dashboard for upvote speed charts, proximity comparisons, and more analytics.
Ben McKay