For survey, reality capture and AEC teams

Turn your point cloud into a realistic 3D scene.

Upload your scan to e57.gs. Explore the result in your browser and share it with a link on view.gs.

We remove people and noise, identify floors and rooms, and create meshes and Gaussian splats you can explore or download.

E57 · LAS · LAZ · PLYPeople and noise removedFloors and rooms foundPay only when it succeeds
Input
E57, LAS, LAZ, PLY or a drone survey
Scanners
Terrestrial, handheld SLAM, mobile, drone
Output
Splat scene, mesh, clean point cloud
Scale
One room to a whole district
Billing
Credits, charged after success

What it does

Everything between the raw scan and a scene your client can use.

The upload runs through a full processing pipeline, not a format converter. Each step is automatic, and each one is something a technician would otherwise do by hand.

Clean scenes

People, carts and ghosts removed before anyone sees them.

A busy site leaves smeared people, parked trolleys and half-closed doors in a scan. e57.gs compares what every scanner station saw: anything one station looked straight through is a passing object, not part of the building. The photographs back this up, catching people the laser missed, and stray points floating in the air are filtered without eating thin, sparse surfaces such as water.

Moving peoplefound from station line-of-sight
People in photosmasked in every panorama
Floaters and edge noiseremoved point by point
Holes in sparse areasfilled from neighbouring surfaces
Building layout

Floors and rooms, found for you.

Storeys are split by height and the stations on them, then each floor is divided into rooms and spaces. Viewers jump between floors and step from station to station.

  • Floor selector and room outlines
  • Station tour with camera positions
  • A point cloud for every floor
Semantics

A scene that knows what it contains.

Every part of the scene carries a label: walls, floors, ceilings, doors, furniture, equipment, people. Labels drive the clean-up and travel with the model for search and filtering.

Meshes

Textured meshes from the same capture.

Alongside the splat, get a textured mesh built from the scan geometry and coloured from the scanner's photographs, ready for CAD review, game engines or 3D printing.

  • GLB and glTF
  • OBJ and STL on request
Any capture

Terrestrial, handheld, drone or aerial.

Static scanner setups, SLAM backpacks and handhelds, Riegl drone surveys with GNSS, and merged clouds with no station data. Sparse aerial clouds get coarser, gap-free surfaces; dense rooms keep 8 mm detail.

Scale

Whole buildings stream on a phone.

Large captures are split by floor and room, processed in parallel and merged into one model. A streamed format loads the area in view first, so a multi-storey building opens in a mobile browser.

Accuracy

Built on your measurements.

The scene starts from your aligned geometry. On the Best level, photographs are aligned to it and station poses refined before training, so colour lands on the right surface.

Product demo

From E57 upload to an interactive Gaussian Splat.

Create Gaussian splats from E57 scans. Upload your scan, follow its progress, inspect the completed scene in the browser and download the result from the same workspace.

e57.gs workflow showing an E57 upload, conversion progress and an interactive Gaussian Splat preview in the browser
Upload E57 → process in the cloud → preview the Gaussian Splat in your browser.Demo input: an E57 file exported from Matterport.

How it works

Three steps, and only one of them is yours.

Step 01

Upload the scan

Straight from the scanner or your registration software. We check the file first and tell you which processing it supports before a credit is spent.

Step 02

We process it

Clean-up, floors and rooms, labels, then the scene itself, on cloud CPUs and GPUs. Follow every stage live in your workspace.

Step 03

Share and download

Send a link anyone can open without an account, and download the splat, the mesh and the point clouds for your own tools.

Inputs

Bring the file your workflow already produces.

No manual clean-up pass before upload. Scans with panoramic imagery and station positions get the full photoreal path; scans without them still get a clean, coloured scene.

CaptureFormatsTested withPhotoreal training
Terrestrial laser scanner
stations with panoramas
E57
Leica RTC360, Leica BLK360, FARO Focus, Trimble X7, Matterport Pro3Yes
Handheld and mobile SLAM
trajectory, optional cameras
E57LASLAZPLY
NavVis VLX, MLX and C-series, FJD, Manifold, GreenValley LiGrip, CHCNAV, FeimaWith camera images
Drone survey
laser plus photographs
RieglLASLAZ
Riegl with GNSS-tagged imageryYes
Merged or aerial cloud
no station data
E57LASPLY
City-scale aerial LiDAR, exported merged cloudsScan colour only

Outputs

One upload, every deliverable.

Everything below comes from the same job and lines up in the same coordinates as your scan.

LINK

Shareable browser viewer

Opens on any laptop or phone, with floor switching and the station tour. No install, no account for viewers.

SOG

Compressed and streamed splats

A compact SOG for small scenes and a streamed, level-of-detail SOG for whole buildings.

PLY

Gaussian PLY and SPLAT

Full Gaussian attributes for editing tools and game engines, or the compact SPLAT format for viewers.

GLB

Textured mesh

Geometry from the scan, texture from the photographs, in GLB or glTF.

PLY

Clean point clouds

The cleaned cloud per floor, with people and noise already removed.

JSON

Floors, rooms and tour

Storey heights, room outlines and station positions, ready for your own viewer or plans.

Quality levels

Fast preview or full photoreal. You choose per job.

GoodBest
What you getA clean scene coloured from the scan, every surface sized to the scan's densityPhotoreal scene trained on the scanner's photographs, with view-dependent light
Clean-up, floors, rooms, labelsIncludedIncluded
Photo alignment and pose refinementNoIncluded
Mesh and point cloudsIncludedIncluded
Credits per room1 credit2 credits
Best forQuick review, sparse or photo-less scans, large areasClient walkthroughs, marketing, remote site visits

Output inspection

How to verify your converted result.

Use the browser preview to examine appearance between viewpoints, then download the Gaussian PLY and inspect the attributes written by the converter.

Degree-3 spherical harmonics

A full degree-3 schema contains 45 f_rest_* coefficients beyond the base color. The file exposes them for independent inspection.

View-dependent appearance

Orbit a glossy floor, window or polished surface and check whether highlights respond naturally as the camera moves.

Densification and pruning

The trained Gaussian count can differ from the input point count because optimization adds useful detail and removes low-value primitives.

Measured limitations

Missing views, glass, mirrors, moving objects, poor registration and sparse imagery can still create gaps or artifacts. Every result should be reviewed.

Related workflows

More ways to work with scan data.

Pricing

Credits per room, charged on success.

No subscription, no seat licences and no minimum. Credits do not expire, and a job that fails is never charged.

Welcome bonus

Start free

Standard Starter pack$15 · 3 credits

15creditsFREE

For every new user. No purchase required.

  • Browser viewer
  • Shareable results
  • One-time welcome bonus
Get free credits

Automatically added after your first successful upload.

Studio

$200

50 credits

  • 50 credits
  • Browser viewer
  • Shareable results
  • Charged only after success
Create account

Questions

Straight answers.

Which files can I upload?

E57 from terrestrial and mobile scanners, LAS and LAZ, PLY point clouds, and Riegl drone surveys. Scans with panoramic imagery and station positions get photoreal training; scans without them get a clean scene coloured from the points.

How are people and moving objects removed?

Where a scanner station looked straight through a spot that another station recorded as solid, that spot was a passing object, so it is removed. The scanner's photographs are checked as well, which catches people the laser only grazed. You can always review the result before sharing it.

Does it work on buildings with several floors?

Yes. Floors are split automatically and each floor is divided into rooms. Large buildings are processed floor by floor and merged into one model that streams in the browser, with a floor selector for viewers.

What are the semantic labels for?

They tell the pipeline what each part of the scene is, so it can treat people, glass and structure differently. They also travel with the model, so you can filter or search the scene by class.

Can I get a mesh instead of a splat?

Every job can also produce a textured mesh from the same capture, in GLB or glTF. OBJ and STL are available for CAD and printing workflows.

How long does it take?

A room-scale scan is usually ready in under an hour. Whole buildings take longer, scaling with the area and the number of floors; you can follow every stage live.

Do I need credits before I upload?

Uploading is free. After floor analysis, the selected rooms and quality determine the credits needed to start processing. Good costs 1 credit per room and Best costs 2. Credits are charged only after a successful result; failed jobs are never charged.

How long are my files kept?

e57.gs is processing, not permanent storage. Retention follows the limits shown in the service, so keep your own copies of source files and outputs. Share links can be opened by anyone who has the link.

Is e57.gs a converter or a training service?

It is an online E57 to Gaussian Splat converter whose full-training path runs 3D Gaussian Splatting optimization. It does more than map each input point to one fixed-color Gaussian: compatible scanner imagery and poses provide observed views for novel view synthesis.

How can I inspect the result's appearance data?

After a successful conversion, inspect the downloaded Gaussian PLY header. A degree-3 result contains 45 f_rest_* properties in addition to the base color coefficients. This confirms the expected field structure, but not optimization by itself. You can also orbit reflective or glossy areas in the browser and look for appearance that changes with the viewpoint.

What happens if an E57 has no usable imagery or camera poses?

Geometry and per-point RGB can still produce a viewer-ready geometry-based splat, but fixed appearance is not the same as full novel-view training. It cannot recreate reflections, hidden surfaces or photographic observations that were never captured.

What output formats are available?

Successful jobs expose the Gaussian Splat assets produced by the converter, commonly PLY and/or SPLAT depending on the completed result. The Gaussian PLY can preserve optimized appearance attributes for compatible tools.

What is the difference between PLY and SPLAT?

PLY is a flexible container that can store point or Gaussian attributes and is useful for compatible editing and interchange tools. SPLAT is a compact viewer-oriented representation; feature support varies by viewer.

Why does training take longer than direct format conversion?

Direct point remapping is mainly bounded by file parsing and disk I/O. Full 3DGS training runs iterative GPU optimization, adaptive densification and pruning. A typical full-training job takes about 12–40 minutes, depending on the source data and available compute.

Your next site visit could ship as a walkthrough.

Create an account, upload one scan and see the clean, navigable result before you decide anything.

Create an account →