Thought leadership
Stop telling. Start showing.
Marlin Prager
July 15, 2026
•
6 min read
A few weeks ago, our Doug Smith built a working interactive 3D product experience in an afternoon. He started with a 3D model, uploaded it to Miris, pointed an AI coding tool at our SDK documentation, and deployed a web page the same day that showed what 3D brings to a retail product detail page. He then walked into a partner meeting and helped the team across the table see what was possible, with no engineering involvement, no scoping call, and no weeks of runway.
With finite engineering time, committing weeks to a prototype means spending real money on activity that competes with active product priorities and customer commitments. Most ideas never clear that bar, and the technical work either lags behind the commercial conversation or, more often, never starts.
Generative AI has helped to collapse the loop, and at Miris we are taking full advantage of it. As the infrastructure for 3D delivery, we want to help teams envision what is possible when the four bottlenecks of 3D delivery (speed, fidelity, scale, and cost) stop setting the ceiling.
Two things have historically stood between a company and a production-quality 3D experience.
The first was delivery economics. High-fidelity 3D and cost-effective delivery were, for a long time, mutually exclusive. The standard approach, streaming rendered video frames from a cloud GPU per viewer, meant costs that scaled directly with traffic. Every concurrent user required dedicated server resources. The economics only worked for low-traffic, high-margin situations, and at any real scale they collapsed.
Miris addresses this structurally. Assets are conditioned once upstream, so the expensive optimization step happens at ingest, not at delivery. After that, content streams at CDN economics, with no cloud GPU per viewer and no capacity ceiling. The cost model changes from infrastructure rental to bandwidth.
The second barrier was reach. Even after the delivery economics improved, integration complexity determined where 3D could actually go. Building a usable application still required significant engineering time from individuals with graphics expertise, pipeline knowledge, and performance tuning. That overhead meant 3D stayed in the hands of specialized teams, and the experiences that would have driven the most commercial value, product configurators, sales demos, e-commerce product pages, never got built.
This is where we made a deliberate product decision. Rather than requiring a full 3D toolchain, we kept the integration surface as small as we could: upload an asset, drop in a script tag, embed a viewer. The same surface works for every team, regardless of what they are building. That is how we lower the barrier far enough that teams without dedicated 3D engineering resources can actually ship.
A third barrier is worth naming honestly: the content itself. Building a high-quality 3D asset is real work, and it stays real work. Materials, lighting, and geometry still have to be authored well, and no infrastructure layer changes that. What has historically made it harder is the second half of that job, getting the asset into a format that runs and looks good across every device, which has demanded specialist optimization skills and significant pipeline investment. That is the part Miris changes. The conditioning step handles optimization and delivery, so the constraint that used to shape how assets were built, keeping them light enough to ship, no longer applies. What that unlocks for creative teams is a larger topic, and one we will take up in a follow-up post.
In preparing for a partner conversation about luxury commerce, the question on the table was whether high-fidelity 3D could meaningfully change how an expensive product is presented online. The usual answer would have been to commission a prototype from engineering. Instead, Doug built one himself.
He started with a 3D model of a comparable product, uploaded it to Miris, where it was conditioned once for adaptive streaming, then used the Miris SDK documentation and an AI coding tool to wire together the page. He generated supporting product imagery and deployed the same day.
The result is live. View the prototype.
The streaming infrastructure behind the assets on the page is identical to what a production deployment would use. That is a different kind of sales conversation, and it did not require a graphics engineering team to make it happen.
The difference between telling and showing is dramatic. When you show people what is possible instead of putting up a slide and talking your way through what is possible, you're playing a different game. You connect the dots directly for your customers, instead of hoping they do on their own. — Doug Smith, Miris
The AI coding tool is incidental. What matters is that the Miris SDK was legible and small enough for a non-engineer to wire up a real streaming experience without a graphics background.
This is an intentional product decision, not an accident of simplicity. We have built and iterated on prototypes across multiple surfaces and use cases precisely to pressure-test that integration surface. What you see in the handbag demo is the same SDK anyone gets when they sign up for the Miris beta. The surface does not change based on who you are or what you are building.
When the integration overhead drops far enough, the loop between idea and prototype compresses dramatically. Concepts that previously required scoping calls, engineering allocation, and weeks of calendar time can now be explored in a day. That changes what gets tested, which ideas make it to a partner meeting, and how the conversation goes.
The handbag prototype is one of many we have built. We have been experimenting across a range of surfaces and use cases, and we plan to document more of them across our website. Not as a product showcase, but as an honest account of what is actually possible right now, and where the edges are. One of those pieces will take up the question we set aside earlier: what changes for creative teams when the delivery constraint comes off.
If your team is evaluating 3D for a product surface and wants to see what the integration actually looks like, the Miris beta is free. Upload an asset and see how far you get.
