our Platform

Upload once, stream everywhere

The adaptive 3D streaming platform. Upload your assets, stream to any device, and scale without infrastructure — from prototype to production.
features
From raw asset to streaming endpoint. Automatically.
Miris lives between your 3D content and your users. We generate streaming-optimized assets deliverable to any device. No pipeline to build. No infrastructure to manage. Just upload and deliver.
Upload and process

Upload USD assets up to 10GB. Miris generates streaming-optimized files ready for deployment.

Adaptive streaming

Quality adapts in real time to device and network conditions for smooth performance everywhere.

Analytics

Monitor stream performance: load times, CDN cache hits, sessions, and device breakdown.

Embed SDK

Use the Miris Web SDK to add real-time 3D streaming to any React, Vue, or JS app.

CDN delivery

Assets stream from the nearest edge node. Global delivery with consistent low latency.

3D viewer

Share assets with clients and teammates as quickly as you would share an image.

What if 3D worked like video?
3D on the web has always meant a tradeoff. If you want 3D content fast, you crush the quality. If you need reach, you're paying per user, or praying they wait. Miris changes that. First frame delivered sub-second. No heavy files to ship. Works on any device, everywhere.
Platform

How it works

Stream every asset instantly while keeping your infrastructure simple and scalable.

Start with any OpenUSD object or imagery

Import any 3D asset in OpenUSD format. Miris accepts models up to 10GB with textures, materials, and lighting data intact. Author in your digital content creation tool of choice, whether you’re working with CAD models, photogrammetry scans, or game assets, start with what you have.

Supports textures, materials, and lighting data.
No manual stripping or re-export required.
Up to 10GB per asset.
Large architectural models and high-poly scans accepted.
Export guides for Blender, Maya, Houdini, and more.
Step-by-step USD export documentation.

Build the foundational asset

Miris converts your source file into a streaming-optimized format. Multi-resolution LODs are generated automatically. Geometry is compressed. Textures are optimized. Visual fidelity is preserved from the original. This is the foundation that makes adaptive 
delivery possible.

Automatic LOD generation.
Multiple levels of detail from a single source file.
Lossless visual fidelity.
Source materials, lighting, and textures preserved.
Export guides for Blender, Maya, Houdini, and more.
Step-by-step USD export documentation.

Optimize for every device and network

Miris analyzes each asset’s geometry complexity, texture resolution, and material composition to determine optimal compression and encoding strategies. The result: maximum visual quality delivered within your viewers’ bandwidth and device constraints. Quality adapts per viewport, per device, per network condition.

Per-asset analysis.
Compression strategies based on actual geometry and texture complexity.
Bandwidth-aware encoding.
Optimized for fiber, cellular, and constrained networks.
Device-aware quality tiers.
Desktop GPU, mobile, and XR headsets each get appropriate fidelity.

Your asset is stream-ready

Your processed asset is now optimized for progressive delivery. Content is visible in seconds. Full detail loads as users explore. Quality transitions are automatic, adapting to viewport distance, device capability, and network conditions in real time.

First frame in milliseconds.
Visible before full detail arrives.
Progressive loading.
Detail increases as users zoom and rotate.
Resolution-independent.
Same source, optimized output per device.

Data where you need it

Miris routes your content through a global CDN. Assets are cached at edge nodes close to your users. Predictive loading ensures sub-second response times regardless of asset complexity or user location. No origin server bottleneck. Performance is predictable from day one.

Global CDN delivery.
Nearest edge node serves every request.
Predictive caching.
Sub-second response times worldwide.
No per-session compute.
Costs scale with storage, not concurrency.

Integrate into your workflow

Connect Miris to the tools you already use. Drop a component into React, Vue, or vanilla JS — our Web SDK integrates directly with three.js as the runtime environment. No workflow changes. No re-architecture. Your tools, your pipeline, our delivery layer.

Web SDK.
JavaScript for web with TypeScript types. Integrates with three.js and any WebGL-compatible renderer.
Zero lock-in.
Miris sits alongside your current stack. Standard components, standard patterns. Remove it and your app still builds.
More platforms (comings soon).
Native SDKs for mobile and game engines are on the roadmap. Join the beta to shape what's next.

Build once, stream anywhere

Embed in any web app with a React, Vue, or vanilla JS component. Your users don't download anything — they just open a link and see your model instantly, whether they're on a laptop, phone, or WebXR headset. One asset, one endpoint, every browser.

Web SDK.
React, Vue, vanilla JS. ~18KB gzipped. TypeScript types included.
Cross-device:
Runs in any modern browser — desktop, mobile, or headset. No native app required.
Platform

How it works

Stream every asset instantly while keeping your infrastructure simple and scalable.

Start with any USD object or imagery

Import any 3D asset in OpenUSD format. Miris accepts models up to 10GB with textures, materials, and lighting data intact. Author in your digital content creation tool of choice, whether you’re working with CAD models, photogrammetry scans, or game assets, start with what you have.

Supports textures, materials, and lighting data.
No manual stripping or re-export required.
Up to 10GB per asset.
Large architectural models and high-poly scans accepted.
Export guides for Blender, Maya, Houdini, and more.
Step-by-step USD export documentation.

Build the foundational asset

Miris converts your source file into a streaming-optimized format. Multi-resolution LODs are generated automatically. Geometry is compressed. Textures are optimized. Visual fidelity is preserved from the original. This is the foundation that makes adaptive 
delivery possible.

Automatic LOD generation.
Multiple levels of detail from a single source file.
Lossless visual fidelity.
Source materials, lighting, and textures preserved.
Export guides for Blender, Maya, Houdini, and more.
Step-by-step USD export documentation.

Optimize for every device and network

Miris analyzes each asset’s geometry complexity, texture resolution, and material composition to determine optimal compression and encoding strategies. The result: maximum visual quality delivered within your viewers’ bandwidth and device constraints. Quality adapts per viewport, per device, per network condition.

Per-asset analysis.
Compression strategies based on actual geometry and texture complexity.
Bandwidth-aware encoding.
Optimized for fiber, cellular, and constrained networks.
Device-aware quality tiers.
Desktop GPU, mobile, and XR headsets each get appropriate fidelity.

Your asset is stream-ready

Your processed asset is now optimized for progressive delivery. Content is visible in seconds. Full detail loads as users explore. Quality transitions are automatic, adapting to viewport distance, device capability, and network conditions in real time.

First frame in under 2 seconds.
Visible before full detail arrives.
Progressive loading.
Detail increases as users zoom and rotate.
Resolution-independent.
Same source, optimized output per device.

Data where you need it

Miris routes your content through a global CDN. Assets are cached at edge nodes close to your users. Predictive loading ensures sub-second response times regardless of asset complexity or user location. No origin server bottleneck. Performance is predictable from day one.

Global CDN delivery.
Nearest edge node serves every request.
Predictive caching.
Sub-second response times worldwide.
No per-session compute.
Costs scale with storage, not concurrency.

Integrate into your workflow

Connect Miris to the tools you already use. Drop a component into React, Vue, or vanilla JS — our Web SDK integrates directly with three.js as the runtime environment. No workflow changes. No re-architecture. Your tools, your pipeline, our delivery layer.

Web SDK.
JavaScript for web with TypeScript types. Integrates with three.js and any WebGL-compatible renderer.
Zero lock-in.
Miris sits alongside your current stack. Standard components, standard patterns. Remove it and your app still builds.
More platforms (comings soon).
Native SDKs for mobile and game engines are on the roadmap. Join the beta to shape what's next.

Build once, stream anywhere

Embed in any web app with a React, Vue, or vanilla JS component. Your users don't download anything — they just open a link and see your model instantly, whether they're on a laptop, phone, or WebXR headset. One asset, one endpoint, every browser.

Web SDK.
React, Vue, vanilla JS. ~18KB gzipped. TypeScript types included.
Cross-device:
Runs in any modern browser — desktop, mobile, or headset. No native app required.
Integrations
THE DIFFERENCE
One asset. Every device.
Video streaming solved adaptive delivery for pixels decades ago. Miris applies the same architecture to 3D. Geometry, materials, and lighting stream to any device based on what it can handle — no file downloads, no server-side rendering, 
no quality compromise.
01.
Speed

First frame renders in under 2 seconds. We stream what's visible first, then progressively load detail as users explore.

02.
Fidelity

Assets stream at native resolution. Level-of-detail adapts automatically based on viewport and device capability.

03.
Scale

Upload once, stream anywhere. Cross-platform delivery to phones or workstations — no manual optimization required.

04.
Cost

You pay for what you stream, not for standing up or maintaining rendering infrastructure. No GPU servers.