# Flai > Flai automates point cloud classification and analysis, turning raw LiDAR data into actionable intelligence in minutes instead of days. One AI engine (FlaiCortex) powers nine pretrained models for semantic 3D mapping across aerial, mobile, national, utility, forestry and defence. Flai is 3D spatial intelligence for LiDAR point clouds. Flai builds the best point cloud semantic segmentation algorithm in the world. The FlaiCortex engine turns raw captures into typed 3D objects and surfaces, with typical time savings of 50-99% versus manual editing. It runs on the cloud, on your local NVIDIA GPU, or fully air-gapped on your own hardware. Point clouds in, decisions out. All URLs below use the canonical domain https://www.flai.ai. ## Products - [Flai Cortex CLI](https://www.flai.ai/products/cli): The full engine on your hardware. Local NVIDIA GPU, ~2x faster than cloud, fully offline and air-gap capable, batch processing across GPUs. - [Flai Web App](https://www.flai.ai/products/web-app): Classify in your browser. Upload LiDAR, pick a template, download classified results. No GPU, no install. Free tier of 10,000 processing units per month. - [Flai Open Lidar Hub](https://www.flai.ai/products/open-lidar-hub): An open European 3D baseline. Over 300 TB of open LiDAR across 16 countries, free to browse, download and build on. - [Flai Lib](https://www.flai.ai/products/library): FlaiCortex embedded in your product. On-device and edge inference, fully offline, the same models deployed at the edge. - [Flai Open Source Products](https://www.flai.ai/products/open-source): Free, open tools for LiDAR. The Flai Python SDK (flai-sdk), the Flai QGIS plugin, and the upcoming ESA-funded EO4For forest-inventory portal. ## AI models Nine pretrained models, all powered by the one FlaiCortex engine. - [Aerial mapping](https://www.flai.ai/models/aerial): Flagship airborne-LiDAR model. Full-scene classification from a single capture, proven at national scale. - [Noise filtering](https://www.flai.ai/models/noise): Cross-cutting preprocessing. Detects and removes isolated and building-edge noise ahead of every downstream model. - [Utilities and powerline](https://www.flai.ai/models/powerline): High-detail power corridors resolved down to insulators, guy wires and tension wires for inspection and vegetation management. - [Mobile mapping](https://www.flai.ai/models/mobile): Dense, close-range vehicle LiDAR classified to asset-inventory granularity: road surface, furniture, signs, poles and vehicles. - [Defence and public safety](https://www.flai.ai/models/defence): Terrain and typed objects for GEOINT, OCOKA and JIPOE workflows. Deployable fully air-gapped. - [Bathymetry](https://www.flai.ai/models/bathymetry): Through-water mapping. Separates water surface, column and seabed, and types vessels, piers, cranes, buoys and breakwaters. - [Forestry](https://www.flai.ai/models/forestry): Forest structure resolved into ground, vegetation strata and individual tree stems for inventory, canopy-height models and biomass. - [Railroad](https://www.flai.ai/models/railroad): Rail corridor classified into track bed, rails, the full catenary system, masts and trackside infrastructure. - [Indoor / BIM](https://www.flai.ai/models/indoor): Industrial-facility scans classified into piping, HVAC, cable trays, beams, gratings and supports for scan-to-BIM. ## Classification categories Every class Flai can segment, grouped by model. Class sets are configurable and extendable per deployment via retraining. Aerial mapping: Other (man-made), Ground, Vegetation, Buildings (roof), Low noise, Water, Bridge decks, High noise, Roof objects, Vehicles, Walls / facades, Points below ground, Low-voltage wires, High-voltage wires, Railroad wires, Low-voltage towers, High-voltage towers, Railroad towers, Fences, Impervious surfaces, Large vehicle, Containers, Street lamps, Roof solar panels, Solar panels, Wind turbines, Antennas. Noise filtering: Low isolated noise (below ground), High isolated noise (dust, birds, cloud), High building noise (interiors / edges), Points below ground. Utilities and powerline: Other, Ground, Vegetation, Buildings, Low noise, High noise, Vehicles, Low-voltage wires, High-voltage wires, Railroad wires, Low-voltage towers, High-voltage towers, Railroad towers, Fences, Insulators, Guy wires, Tension wire. Mobile mapping: Other, Road, Sidewalk, Other ground, Traffic island, Buildings, Trees, Other vegetation, Traffic lights, Traffic signs, Wires, Masts, Pedestrians, Two-wheel, Mobile four-wheel, Stationary four-wheel, Noise, Tree trunks. Defence and public safety: Ground / terrain, Vegetation, Buildings, Infrastructure, Tree trunks, Fallen trees / debris, Vehicles, Heavy vehicles / plant, Equipment / installations, Aircraft / rotary-wing. (Classes are configured per deployment via sovereign, in-network retraining.) Bathymetry: Other, Ground, Vegetation, Building, Low noise (water column), Water, High noise, Vehicles, Containers, Bathy ground (seabed), Naval mines, Pier, Vessel, Buoy, Crane, Breakwater, Underwater structures. Forestry: Unclassified, Ground, High vegetation / canopy, Noise, Tree trunks / stems. (Species and biomass heads extend the class set via retraining.) Railroad: Other, Noise, Terrain, Tunnels, Bridges, Buildings, Electric boxes, Manholes, Fences, Rail, Masts, Isolators, Traffic signs, Other signs, Cantilever, Powerline, Catenary wire, Droppers, Support wires, Return wire, Cable channel. Indoor / BIM: Background, Beams, Cable trays, Civils, Gratings, Guardrails, HVAC, Ladders, Piping, Supports. ## Industries - [Topographic Mapping](https://www.flai.ai/wide-area-mapping): Wide-area and topographic object extraction from airborne LiDAR. - [Mobile Mapping](https://www.flai.ai/mobile-mapping): Street-level asset inventory and city modelling from vehicle LiDAR. - [BIM and Indoor Mapping](https://www.flai.ai/bim-indoor-mapping): Scan-to-BIM-ready typed objects from indoor and terrestrial scans. - [Forestry](https://www.flai.ai/forestry): Single-tree inventory, canopy-height models, biomass and carbon stock. - [Utilities and Powerlines](https://www.flai.ai/power-lines): Vegetation management, wire vectorization and clearance analysis. - [Mining](https://www.flai.ai/mining): Classification and analysis for mine sites and volumes. - [Defence and Security](https://www.flai.ai/defence-and-security): 3D situational awareness, deployable in secure or restricted networks. - [Bathymetric Mapping](https://www.flai.ai/bathymetric-mapping): Through-water seabed and maritime-infrastructure mapping. - [Computer Vision R&D](https://www.flai.ai/computer-vision-rd): Production 3D segmentation, labelled data and a foundation engine to build on. - [Robotics and Navigation](https://www.flai.ai/robotics-navigation): On-device semantic 3D perception for robots and autonomous platforms. ## Capabilities - [Pretrained AI Models](https://www.flai.ai/models): Catalog of the nine ready-to-run models. - [Custom AI Models](https://www.flai.ai/solutions/custom-ai): Train and adapt models on your own data. - [Classification](https://www.flai.ai/solutions/classification): Automated point cloud classification. - [Vectorization](https://www.flai.ai/solutions/vector-extraction): Extract typed vectors from classified clouds. - [Rasterization](https://www.flai.ai/solutions/rasters): Generate DTM, DSM and derived rasters. - [Ontology](https://www.flai.ai/solutions/ontology): The class ontology behind the models. - [On-the-fly Maps](https://www.flai.ai/solutions/on-the-fly-maps): Maps generated directly from classified data. - [Forestry products](https://www.flai.ai/solutions/forestry-products): Forest-inventory outputs and metrics. ## Key pages - [Pricing](https://www.flai.ai/pricing): Volume-based licensing and pay-per-download. - [Platform overview](https://www.flai.ai/platform): Processing flows, 3D viewer, annotation tools, AI learning point and self-hosting. - [Contact](https://www.flai.ai/contact): Book a demo or request a trial. ## Optional - [Blog](https://www.flai.ai/resources/blog): Articles and updates. - [News](https://www.flai.ai/news): Programme selections, partnerships, certifications and awards. - [Templates](https://www.flai.ai/resources/templates): Reusable processing-flow templates. - [Webinars](https://www.flai.ai/webinars): Recorded and upcoming sessions. - [Company](https://www.flai.ai/company): About Flai. - [Documentation](https://docs.flai.ai): Product and API docs. - [Web app login](https://app.flai.ai): The Flai cloud application.