The beta is live

Teams can now produce high-fidelity, photorealistic 3D faster than ever. But getting it to users at scale is still the bottleneck.

That gap has forced every team building with 3D into the same tradeoff: sacrifice fidelity for reach, or sacrifice cost for quality. We built Miris to eliminate that tradeoff. In February, we announced the Public Beta and explained why adaptive spatial streaming changes the economics of 3D delivery.

Today, the platform is live. Teams can sign up, upload assets, and start streaming.
‍

‍

This post is about what you can do with Miris right now, and what the beta experience looks like in practice.

‍

What you get today

When you sign up for the beta, you get immediate access to:
‍

• Sample assets to explore before uploading your own. We've included a set of sample 3D assets so you can see Miris streaming in action within minutes of signing up. No asset preparation required to evaluate the platform.
  ‍
• Miris web SDK. The beta ships with a web SDK that lets you stream adaptive, high-fidelity 3D directly in a browser. Upload an OpenUSD asset, integrate a few lines of code, and start streaming to any device. The SDK includes Three.js support for teams already building with it, but a preexisting Three.js application is not required. Unity, Unreal Engine, and NVIDIA Isaac Sim integrations are on the roadmap.
  ‍
• Self-serve developer documentation. The entire onboarding flow is self-serve. Our documentation walks you through asset upload, SDK integration, and streaming configuration. No sales calls required to get started.
  ‍
• Free access during beta. During public beta, you get unlimited uploads, 5 high-fidelity processing credits per month, 1000 views per account per month before watermarking, and 10GB maximum file size per asset. The goal is simple: go from signup to streaming your own 3D asset in hours.
  ‍

‍

What teams are already building

We're not launching into a vacuum. Early partners are already integrating Miris into production workflows.

‍

‍

Playbook, the creative asset management platform, plans to use Miris to provide users with fast, high-fidelity 3D asset review directly in the browser. Instead of downloading multi-gigabyte files to inspect a model, their users will stream and interact with assets at source quality.

Voxel51, the open-source toolkit for visual AI, is integrating Miris for robotics and industrial applications. Their teams work with large-scale 3D datasets captured from real-world driving environments. Miris lets them stream those datasets to any workstation without requiring massive local downloads or dedicated GPU infrastructure for each viewer.

These are the kinds of workflows Miris was built for: teams that already have high-fidelity 3D content and need a practical way to deliver it.

How Miris works (quick recap)

Miris is an adaptive spatial streaming platform. You upload 3D source assets (currently OpenUSD, with additional formats supported over time). At ingest, Miris processes the asset and generates an optimized version for streaming. There is no manual optimization step; Miris handles it.

When a user loads the experience, the scene begins rendering immediately. Only the data required for the current view is delivered first. Additional detail streams adaptively based on network conditions, device capabilities, and viewing behavior. A 1GB source asset loads and becomes interactive as quickly as a simple 5MB glTF. 

The key architectural distinction: Miris streams spatial data that is reconstructed on the client device. There are no cloud GPU dependencies at the point of delivery. Costs are similar to streaming video, rather than running cloud GPUs.

The result is sub-second time-to-interaction for complex assets, high fidelity without massive downloads, and infrastructure that scales without per-user cloud GPU costs.

‍

‍

Why this matters now

Three things have converged to make this the right moment for developers to evaluate Miris.

• 3D content creation has outpaced distribution. Generative AI, improved photogrammetry, and broader OpenUSD adoption have made it faster and cheaper to produce high-fidelity 3D. But the delivery bottleneck remains. Teams are producing assets they cannot practically distribute at scale. The creation-to-delivery gap is widening.
  ‍
• The cost of workarounds is increasing. Every team that builds custom optimization pipelines, maintains platform-specific asset variants, or manages pixel streaming infrastructure is spending engineering time on delivery rather than on product differentiation. The longer you defer, the more that costs compound.
  ‍
• The beta is the fastest way to validate. Miris is not a slide deck or a waitlist. The platform works end-to-end today, from asset upload to SDK integration to live streaming. Developers in the beta can test the full pipeline against their own content, measure load times, evaluate fidelity, and see how streaming behaves across devices.

‍

Get started today

Sign up for the free Public Beta. Access is limited. We're onboarding developers in waves to ensure quality support, so sign up today to lock in your spot. The beta runs through summer, ahead of general availability later this year.

If you're building web experiences with 3D (product configurators, architectural visualization, digital twins, simulation, interactive retail) and you've been working around the limitations of current delivery methods, this is the moment to try a different approach.

We're building Miris with developer feedback. What you tell us during beta shapes what we ship.