2.4 Software Map
Four applications. Two pipelines. Each has a specific, non-interchangeable role.
The Four Applications
The XGRIDS software ecosystem consists of two field apps (one per device family) and two desktop processing applications (one per output pipeline). None of these are interchangeable. Using the wrong app with a device results in a connection failure. Processing data through the wrong desktop application produces either an error or an output that does not match the project requirement.
LixelGO
iOS and Android. Connects to the L3, L2 Pro, and K2 via WiFi. Controls scanning start and stop, displays a real-time point cloud preview during the session, manages control point marking, and configures RTK and NTRIP settings. This is the operator's primary field interface for all three LiDAR devices. The L3 can also run a scan from its built-in display with no phone connected. It does not process data.
Android: 8 GB RAM, Snapdragon 865 or equivalent, Android 10 or later
iOS: A12 Bionic or later, 4 GB or more RAM, iOS 17.5 or later recommended
Both: Bluetooth 5.0 or later, WiFi 5 or later, personal hotspot capability required for RTK/NTRIP
LCC Scan
iOS and Android. The PortalCam's dedicated field control app. Not interchangeable with LixelGO. Controls PortalCam scanning, displays a real-time quality preview, manages control point marking during the session, and monitors device status. If you own a PortalCam, this is your field app. If you own an L3, L2 Pro, or K2, LixelGO is your field app.
Direct Connect via phone hotspot: most reliable mode, recommended for all scanning sessions
Bridge mode via shared WiFi network: available but more susceptible to interference
LixelStudio
Windows 10/11 desktop application. Processes L3, L2 Pro, and K2 raw data into point clouds. Handles SLAM trajectory optimization, RTK and PPK coordinate transformation, ground control point application, Map Fusion for multi-session projects, and export to LAS, LAZ, PLY, E57, OBJ (with texture), and RCP. AMD GPUs are not supported. Requires an NVIDIA GPU.
CPU: Intel i7 11th generation. Recommended: Intel i9 12th generation
GPU: NVIDIA RTX 3060 or higher. Recommended: NVIDIA RTX 3070 or higher. AMD not supported
RAM: 64 GB, at both the minimum and the recommended configuration
Storage: 1 TB minimum, 1 TB SSD recommended. Install path must not contain spaces or non-ASCII characters
OS: Windows 10 or Windows 11
LCC Studio
Windows 10/11 desktop application. Processes L3, L2 Pro, K2, and PortalCam data into 3D Gaussian Splat models. The Free tier supports single-scene generation, viewing, editing (clip, crop, measure, annotate), asset overlay, publishing, cloud data management, recording and flythroughs, and export to .lcc and .lcc2. The Premium tier adds Map Fusion, Aerial Reconstruction, Ground and Aerial Fusion, HD Enhancement, and Spatial Recognition, plus export to .ply, .usd, and 3D Tiles. Current pricing is on the XGRIDS Store. Requires an NVIDIA GPU with CUDA. AMD GPUs are completely unsupported with no available workaround. AMD CPUs (e.g., Ryzen 9) are supported and recommended by XGRIDS for processing workloads.
.lcc and .lcc2 (Free); .ply, .usd, .usdz*, 3D Tiles* (Premium); .obj or .ply mesh (no texture)
*USDZ requires firmware ≥ 3.0 and Single Model reconstruction. 3D Tiles limited to models with ≤ 4 million Gaussian points.
CPU: Intel Core i7 10700K or AMD Ryzen 7 3700X (8-core desktop). Recommended: AMD Ryzen 9 7950X (16-core desktop)
GPU: NVIDIA RTX 3060 12 GB or NVIDIA Quadro RTX A4000 16 GB. Recommended: NVIDIA GeForce RTX 4090 24 GB, Quadro RTX 6000 Ada 48 GB, or Tesla L40 48 GB. AMD not supported, CUDA mandatory
RAM: 64 GB minimum
NVIDIA driver 581.90 or later
OS: Windows 10 or Windows 11
Device and Software Pairing
Every device has a specific field app and a specific set of processing options. The PortalCam's path is entirely separate from the L3, L2 Pro, and K2 path, from the first moment of connection through the final processed output.
XGRIDS lists the K2 for Single Model reconstruction, Map Fusion, and Aerial-Ground Map Fusion in LCC Studio, and lists the Lixel K series for HD Enhancement. Those fusion and enhancement modes need Premium.
Processing Hardware: What Matters and Why
Memory Is the Scan-Length Ceiling in LCC Studio
LCC Studio holds the working reconstruction in memory, so installed memory sets the ceiling on how much capture one reconstruction can take. Attempting to reconstruct more than your memory supports will cause a mid-run failure. LixelStudio specifies 64 GB at both its minimum and recommended configurations.
Projects requiring scans longer than your RAM supports must be broken into shorter segments and joined using Map Fusion. See Module 5: Map Fusion for the complete planning and processing workflow.
LixelStudio's published graphics floor is NVIDIA RTX 3060 or higher, with NVIDIA RTX 3070 or higher recommended.
LCC Studio: GPU Is the Constraint
LCC Studio is a GPU-dominant workload. CUDA is mandatory. AMD GPUs are completely unsupported with no workaround. AMD CPUs are fully supported, and XGRIDS recommends the AMD Ryzen 9 9950X or equivalent 16-core processor for Map Fusion and Aerial-Ground Fusion workloads. Processing speed and maximum output quality are both determined by your GPU's VRAM.
For Map Fusion and Aerial-Ground Fusion projects in LCC Studio, XGRIDS recommends 16 or more cores, such as AMD Ryzen 9 9950X, 64 GB DDR5 of memory (128 GB for large models of 150 minutes or more at high quality), and NVIDIA RTX 3090 or better.
Large models need more memory. LCC Studio may need 128 GB for large models of 150 minutes or more at high quality. If RAM is insufficient, use Standard or Medium Quality processing settings to reduce the risk of failure. High or Ultra quality settings on underpowered hardware will likely fail mid-process.
Mobile Device Requirements
LixelGO and LCC Scan both run on iOS and Android, to the minimum phone specifications in Section 1. The connection between the field app and the scanner is WiFi.
- Personal hotspot capability is required when using RTK with an NTRIP correction service
- Charge or carry a power bank for sessions exceeding 60 minutes. The field app must remain active throughout the scan
Processing Time Expectations
Running either software on underpowered hardware extends these estimates significantly and increases the risk of failure.
These ratios are based on recommended hardware configurations. Higher quality settings in LCC Studio extend processing time beyond these estimates. Draft quality is significantly faster and appropriate for preview purposes.
LCC Studio's first launch after installation compiles CUDA kernels, which takes 2 to 5 minutes. This is a one-time process per installation.
©2026 Alpine Reality Capture LLC • XGRIDS Pro Guide™ • Site Disclaimer

