Precision Point Cloud Classification & 3D Extraction — Right in Your Browser.

The zero-friction, self-hosted web tool for GIS analysts and surveyors. Clean manual classification, automated 3D object generation, and custom ML model training—all running in Docker.

Docker ReadyFree for Commercial UseZero Cloud DependencyWebGL Accelerated

Fast Manual Edits

Surgical Precision

Rectangle and lasso for fast selection, polygon for granular control, brush for irregular patterns, and hull select to capture whole 3D objects from a few corner points — plus smart selection that identifies objects automatically. Combine tools in multi-step selections and refine any selection by classification.

Rich Visualization

Every Angle on Your Data

Flip a multi-million point survey between classification, elevation, and intensity shading in real time.

Train & Infer

Custom AI Models

Train custom point classification models on your local hardware using your own annotated ground truth.

Re-imagined Manual Classification

Effortless Cleanup of Massive Point Clouds.

Our browser-based editor delivers fluid 60fps interaction with multi-million point datasets. Select, filter, reclassify, and clean up vegetation noise or structural misclassifications with instant visual feedback.

Automated 3D Object Creation

From Point Clouds to CAD-Ready 3D Geometry.

Convert classified LiDAR points into lightweight 3D geometries. Automatically generate roof structures, building envelopes, terrain contours, and street furniture objects. Export seamlessly to standard GIS and CAD formats.

Video loop · placeholderBuilding points highlighting, thresholding, and generating clean 3D OBJ/GeoJSON geometries automatically.

In-Browser Model Training & Inference

Train Classification Models on Your Own Datasets.

No cloud subscription required. Define classes, train custom machine learning segmentation models directly on host hardware, and execute inference on newly acquired point clouds with a clean, low-code interface.

GIF loop · placeholderTraining configuration, epoch loss charts, and immediate inference results overlaying a point cloud.

Under the Hood

Streams Millions of Points. Edits Every Single One.

Built on COPC

Every point cloud you import — LAS, LAZ, or PLY — is converted internally to COPC (Cloud Optimized Point Cloud), an open format that stores points in a compressed octree. The viewer streams exactly the level of detail your current view needs via HTTP range requests: huge surveys open instantly, navigation stays fluid, and no full file ever has to be downloaded.

Voxel-Complete Editing

Streaming brings one editing challenge: at any moment, not every point of an area is displayed. So that edits still cover everything, your first selection builds voxels of configurable size around the selected points — and LiDARic then downloads all remaining points inside that volume. Reclassification applies to every point in the area, not just the ones on screen.

Up and Running in 60 Seconds.

Deploys as a single, lightweight Docker container. Keep all sensitive spatial data on your local network or private server.

bash
# Pull and run the container locally
docker run -d \
  --name lidaric \
  -p 8000:8000 \
  -v lidaric-data:/data \
  lidaric/lidaric:latest
# then open http://localhost:8000
Read Full Deployment Guide →

Ready to Elevate Your Spatial Workflow?

Free for personal and commercial use. Deploy on your hardware today.

Deploy Container Now