Aholo Viewer - An open-source 3D Gaussian browser from Quncore Technology.

What is Aholo Viewer?

Aholo Viewer is an open-source, high-performance web-based rendering engine developed by Quncore Technology, based on 3D Gaussian sputtering (3DGS). The tool employs chunk-level LOD streaming technology, enabling smooth transmission of 3D content across different devices like video. Mobile devices can run massive city-scale scenes with up to 1 billion Gaussian points. The product is compatible with mainstream 3DGS formats such as ply, spz, sog, splat, lcc, and ksplat, and comes with built-in tools for LOD generation, format conversion, and voxel collision volume generation, allowing developers to use it out of the box. It also comes with the Aholo Spatial Intelligence Platform API, covering capabilities such as spatial reconstruction, cloud rendering, and 3D AI model generation, driving the 3D internet from "viewable" to "usable."

Main functions of Aholo Viewer

• Smooth rendering of billion-level Gaussian scenesThrough chunk-level LOD and streaming scheduling, it supports loading and smooth roaming of scenes with 1 billion Gaussian points in the browser in seconds, and the first screen can be entered within 10 seconds.
• Multi-terminal access without installationFully compatible with mobile phones, computers, and VR devices, no client download required, simply open your browser to experience a vast 3D world.
• Multi-format compatibilitySupports mainstream 3D Gaussian formats such as ply, spz, sog, splat, lcc, and ksplat.
• Complete toolchainBuilt-in tools for high-quality LOD generation, data format conversion, 3D Gaussian collider generation, editing and picking.
• Physics collision systemGenerates voxel colliders from reconstructed assets in 3DGS, supports real-time collision queries for rays, capsules, ground detection, and wall obstruction, and supports walking mode and camera obstacle avoidance.
• Cloud Hybrid RenderingBased on OpenUSD, 3DGS and high-fidelity mesh can coexist and be mixed and rendered in the same scene and frame, with real-time streaming in the cloud, allowing for smooth preview even on low-spec terminals.
• Multi-level rendering configurationIt offers three rendering paths: effect priority, performance priority, and extreme performance, balancing image quality and performance as needed.

The technical principles of Aholo Viewer

• Chunk-based LOD TreeThe original 3DGS data is divided into multiple chunks, and different levels of LODs are generated for each chunk. At runtime, the level is switched on a chunk-by-chunk basis. Compared with Spark 2.0's splat-based LOD, the chunk granularity is coarser, the cache hit rate is better, and the scalability is stronger, making it easier to splice and update large-scale scenes at the city level by chunk.
• Streaming batch pullData is loaded according to view frustum priority, controlling memory and video memory usage to achieve efficient scheduling in large scenes.
• Rendering pipeline optimizationIt reduces GPU memory usage through multi-precision data structures, compresses GPU overhead per frame by caching pre-computation and on-demand pass, and improves data access efficiency by combining Morton Sort and detail culling.
• 3D Gaussian SplattingThe scene is modeled as billions of Gaussian ellipsoids with position, shape, color and transparency, and real-time rendering is achieved through differentiable rasterization.

How to use Aholo Viewer