Gaussian splatting turns captured real-world environments into highly realistic, navigable 3D representations. When paired with LiDAR, those same environments can also provide the geometric data needed for measurements and other professional workflows.
At Mirror Labs, we are focused on what happens after the capture. Our web-based platform gives professionals a place to work directly inside these 3D environments by taking measurements, adding annotations, collaborating with others, and reviewing spatial data remotely. Our goal is to help move Gaussian splatting beyond visualization and into practical industry use.
An Insurance Adjuster Use Case
We recently worked with an independent insurance adjuster who saw immediate value in using Gaussian splatting and mobile SLAM capture as part of his claim documentation workflow.
One of the biggest benefits for him was reducing the amount of time he personally needs to spend at a claim site.
Insurance claims often involve damaged environments, including water intrusion, fire damage, structural issues, debris, and other conditions where minimizing unnecessary time on site can improve both efficiency and worker safety.
Traditionally, an adjuster may spend significant time manually documenting a property, taking measurements, collecting photographs, and making sure enough information has been captured before leaving. Terrestrial laser scanning can provide highly detailed spatial data, but depending on the workflow, it can also require more equipment setup, capture time, and specialized expertise.
Mobile SLAM devices such as the XGRIDS K2 create another option.
An operator can move through the property relatively quickly while capturing both the visual and spatial information needed to create a detailed digital representation of the site.
For the adjuster we worked with, that meant spending less time collecting information and more time doing the work that actually requires his expertise, such as reviewing damage, checking measurements, developing an estimate, and making sure the claim is as accurate as possible.
Separating Capture From Expertise
One of the more interesting ideas the adjuster brought up was that he may eventually not need to personally visit every property at all.
Because mobile SLAM systems can be relatively straightforward to operate, he could train another member of his team to perform the capture or use a local capture provider.
The person collecting the data does not necessarily need the same level of insurance expertise as the person evaluating the claim.
Once the property is captured and processed, the adjuster can review the digital environment from his office and perform much of the professional work remotely.
That begins to separate two activities that have historically happened together:
Capturing the physical environment and Applying professional expertise to that environment.
For an independent adjuster, that could mean covering a larger geographic area while spending less time traveling and more time evaluating claims.
More broadly, it raises an interesting question for any field-based profession:
Does the expert always need to be the person holding the capture device?
Why the LiDAR Layer Matters
Photorealistic visualization is incredibly valuable, but for many professional workflows, visualization alone is not enough.
Photographs, video, and 360-degree imagery are excellent for documentation and visual context. Almost anyone can look at them and understand what the environment looked like.
But industries such as insurance, restoration, renovation, and commercial construction often require something more. Decisions may depend on dimensions, quantities, distances, and spatial relationships.
That is where the underlying LiDAR data becomes important.
Devices such as the XGRIDS K2 can produce a realistic Gaussian splat while also capturing an underlying LiDAR point cloud. The Gaussian splat makes the environment intuitive and easy to understand, while the LiDAR provides geometric data that can support measurement-driven workflows.
For an insurance adjuster, those measurements can help support an accurate estimate.
For a contractor or renovation team, they may support estimating, material quantities, planning, or verification.
For commercial construction, the level of accuracy required will depend on the job. Certain workflows may still require survey-grade or more specialized equipment. The important point is that many professional spatial workflows need more than something that simply looks like the real world.
They need a digital representation that can also be used as data.
Bringing Visual and Geometric Workflows Together
This combination also creates an interesting change in how different people can work with the same site.
Historically, teams often worked from two very different types of information.
One group worked from photographs, videos, and 360-degree imagery. These formats were intuitive and useful for understanding what a site looked like.
Another group worked from LiDAR scans, point clouds, terrestrial scanning, surveying data, and other technical spatial datasets. These provided richer geometric information, but often required different software and a more specialized skill set to interpret.
People involved in the same project could effectively be working from two different representations of the same physical environment.
Gaussian splatting with underlying LiDAR begins to bring those worlds together.
Someone who is comfortable with photographs and video can immediately understand the photorealistic Gaussian splat. Someone who needs geometric data can work with the underlying point cloud.
Within Mirror Labs, those users can work from either representation or view them together.
More importantly, they can measure, annotate, communicate, and make decisions within the same 3D environment.
Instead of one person reviewing imagery while another works separately inside a point cloud application, both can reference the same wall, damaged area, structural element, or point of concern.
That starts to make remote collaboration more closely resemble what would happen if both people were physically standing at the job site together, walking through the environment and discussing the same areas.
Where Mirror Labs Fits
There are already plenty of Gaussian splat viewers on the market.
Our goal with Mirror Labs was never simply to build another good way to view a Gaussian splat.
We believe Gaussian splatting has significantly more commercial value than being a highly realistic representation of something that exists in the physical world.
The real opportunity comes from giving that representation the right tools.
Across insurance, restoration, renovation, construction, and other field-based industries, there are real workflows that can benefit from Gaussian splatting. But for that to happen, the technology has to fit into the way professionals already work and provide functionality that helps them complete their jobs.
That is what we have focused on with Mirror Labs.
We want Gaussian splats to become environments people can actually work from, not simply environments they can look at.
The Mirror Labs Workflow
Mirror Labs is entirely web-based and has been designed around making spatial data as easy to work with as possible.
A company can create its own workspace, then organize multiple projects within that workspace. Those projects might represent claims, properties, construction sites, renovation projects, or other physical locations.
Users can be invited into the workspace, and permissions can be managed based on how the organization wants different people to interact with the data.
Once a scan has been captured and processed, it can be brought into the appropriate project.
From there, users can:
Navigate through the Gaussian splat directly in a web browser
View the underlying LiDAR point cloud when available
Take and save measurements
Add annotations directly to locations in 3D space
Tag collaborators on items that require their attention
Notify those collaborators by email
Share projects with other stakeholders
Export relevant project information and documentation
Maintain a persistent digital record of the site
View scan locations geographically across a map
The important part is that the context remains attached to the physical environment.
Instead of sending someone a photograph and trying to explain where an issue is located, a user can identify the exact area in 3D, annotate it, tag the appropriate person, and allow them to review that same location remotely.
Building With Industry Experts
Up to this point, we have intentionally kept access to Mirror Labs relatively controlled.
We have not been focused on getting as many users onto the platform as possible. Instead, we have been working closely with industry professionals to understand how spatial data fits into their existing workflows and what tools actually make that data useful.
The philosophy has been simple:
Do not build features because they are technically possible. Build tools that solve meaningful problems inside real industry workflows.
We are still building alongside those users, but we have reached a point where we are very happy with the platform and are beginning to open conversations with more organizations and professionals interested in using it.
The Bigger Opportunity
We do not think the most important thing about Gaussian splatting is that it creates an impressive 3D visualization.
The larger opportunity is what happens when capturing a physical environment becomes fast enough, accessible enough, and useful enough that everyone involved in the work does not need to physically be there.
Accessible capture hardware can distribute data collection.
LiDAR can provide the geometric information needed for measurement-driven work.
Gaussian splatting can make the environment intuitive and photorealistic.
Mirror Labs can provide the workflow and collaboration layer that allows professionals to actually work from that information.
The result is a much broader workflow:
Capture the physical world once. Bring it online. Preserve the underlying spatial data. Then allow the right people to measure, communicate, collaborate, and make decisions from it wherever they are.