LAStools, the LiDAR toolbox split between two open libraries and a licensed suite

Two open libraries, and a suite that is not open

LAStools is presented as award-winning software for efficient LiDAR processing with LASzip, and the description calls the tools efficient tools for LiDAR processing. The important structural fact is that the project is not one thing.

It consists of LASlib, the low-level processing API, and LASzip, the compression and decompression API. Both parts are free and open source. On top of them sits the LAStools Software Suite, additional tools based on LASlib and LASzip for processing LiDAR data, where some of the tools are free to use and others are licensed.

The open source part lives in the LASlib and LASzip directories in the repository, and the documentation is in the bin directory. The top level also has a CMakeLists.txt, a src/ directory, a LAStools.sln for Visual Studio, a .clang-format file, a COPYING.txt and a LICENSE.txt, and license.html, which together suggest a licensing setup that is not a single standard licence, which is why the repository's own licence field is not classified by GitHub.

The data formats are the other half of the picture. The files are in ASPRS LAS format, versions 1.0 to 1.4, or lossless compressed and otherwise identical twin LAZ format. LAZ is the interesting one, because it is the same data model in a compressed container, and that equivalence is what lets the open compression tools be genuinely useful rather than a niche format.

The code is described as ultra-lightweight, very efficient and super-fast C++, and the tools as highly efficient, multi-core batched, scriptable tools to process LAS, compressed LAZ, Terrasolid BIN, ESRI Shapefiles, ASCII and others.

The last push was on 2026-09-21 and the release tags are v2.0.5 from 2026-05-05, v2.0.4 from 2025-10-24 and v2.0.3 from 2024-04-16.

The open tools, and what each one is for

The open source list is short enough to read, and it covers the operations that come up in almost every LiDAR workflow.

laszip is the powerful, lossless LiDAR compressor that turns large LAS files into much smaller LAZ files. las2las extracts last returns, clips, subsamples, translates and so on, which makes it the general purpose filter. las2txt turns LAS into human-readable and easy-to-parse ASCII, and txt2las converts LiDAR data from ASCII text back to binary LAS, so the two are a round trip. lasmerge merges several LAS or LAZ files into a single LAS or LAZ file, which is the tiling step at the end of a survey. lasindex creates a spatial index LAX file for fast spatial queries, and lasinfo prints out a quick overview of the contents of a LAS file.

Two of the others are worth more attention. lasvalidate determines if LAS files conform to the ASPRS LAS specifications, which is the tool you reach for when a client says the data is broken and you need to know whether it is the data or the reader. And lasdiff compares LiDAR data and reports whether two files are identical or different, which is how you verify that a conversion or a clip did what you expected.

There are also two newer additions. lascopcindex creates a COPC.laz file for a given set of.las or.laz files, which is the cloud-optimised format that lets a reader do range requests against a big point cloud instead of downloading all of it. And e572las extracts the points from the E57 format and stores them as LAS or LAZ files, which is a bridge from the terrestrial scanning format.

Two more sit in the free tier rather than the open tier. lasview visualizes a LAS file with a simple OpenGL viewer, and demzip compresses and uncompresses raster data from ASC, BIL, TIF and IMG format to the compressed RasterLAZ format.

The distinction between the open and free lists is the licensing boundary, and it is drawn in an unusual place. lasprecision appears twice, once in the open source list as analysing the actual precision of the LiDAR points and once in the closed source list as reading LiDAR data in the LAS format and computing statistics about the precision advertised in the header.

The licensed half, where the point limit bites

The closed source list is where the interesting processing lives, and it is where the licensing question comes up.

The rasterising tools are las2dem, which rasters point clouds via a TIN into elevation, slope, intensity and RGB DEMs, las2iso for elevation contours, las2tin for triangulating points into a TIN, and las2shp for turning binary LAS into ESRI's Shapefile format. The classification tools are lasground, which extracts the bare earth by classifying all ground points, lasground_new, an improved version of lasground for complex terrains, and lasclassify, which finds buildings and the vegetation above the ground.

Then there is the cleanup and quality control family. lasduplicate removes duplicate points with identical x and y and z optional. lasnoise flags and removes noise points in LAS, LAZ, BIN and ASCII files. lascontrol does quality checks of elevations against a list of control points, and lasprecision analyses the actual precision of the points. lasoverlap checks the overlap and the vertical and horizontal alignment of flight lines, and lasoverage finds the overage of an airborne collect that got covered by multiple flightlines. lasreturn reports geometric return statistics and repairs the number of returns field based on GPS times.

The geospatial tools are lasdatum, which transforms from one horizontal datum to another, lasvdatum, which transforms LiDAR from ellipsoidal to orthometric elevations using a grid, lasdistance, which classifies, flags or removes points based on distance from polygonal segments, and lasclip, which clips points against building footprints and swath boundaries. lasboundary extracts a boundary polygon enclosing the points, and lasprobe probes the elevation of the LiDAR at a given x and y location.

The practical constraint is stated in the installation section. All licensed tools can be tested for free up to the point limit, roughly 3 to 5 million points. A regional survey is orders of magnitude beyond that, so the point limit is not a generous trial but a genuine boundary between evaluation and purchase, and it is worth knowing before you plan a pipeline around a tool you can only run on a sample.

There is also a BLAST extension, which is the same idea scaled up. blast2dem rasters like las2dem but with streaming TINs for billions of points, and blast2iso does the same for contours.

Installing on Linux, where the library path is the trap

Binary downloads for Windows and Linux are available at rapidlasso.de/downloads, and the LAStools download contains all the binaries of the open source and licensed tools, the LASzip and LASlib libraries, and all the BLAST binaries.

The Windows route is the simple one. All binaries are in the download file, there is a full installer and a binary-only download, and for the binary download you create a directory called c:\lastools, unzip LAStools.zip into it, and run the executables.

The Linux route has a step that catches people out. The first step is to create an installation target and extract the package:

cd ~
mkdir lastools
cd lastools
wget https://downloads.rapidlasso.de/LAStools.tar.gz
tar xvzf LAStools.tar.gz
rm LAStools.tar.gz

Then the dependencies, or an expanded library path to catch the included libraries. The README gives both options. The apt line is:

sudo apt-get install libjpeg62 libpng-dev libtiff-dev libjpeg-dev libz-dev libproj-dev liblzma-dev libjbig-dev libzstd-dev libgeotiff-dev libwebp-dev liblzma-dev libsqlite3-dev

And the alternative is to expand the library path to include the local./lib directory:

export LD_LIBRARY_PATH=./lib:$LD_LIBRARY_PATH

That second option is the reason the bundled libraries exist in the download at all. LASlib and LASzip are open source, and the distribution ships them as shared libraries so the binaries can link against them, which means the binaries need to find them at runtime and the library path has to be set unless you install the system dependencies. The list of packages also tells you what the tools are built against, which is a decent guide to the format support: JPEG, PNG, TIFF, PROJ for projections, LZMA, JBIG2, Zstandard, GeoTIFF, WebP, SQLite and zlib.

Once that is done, you run the executables directly, for example./las2las64.

The examples are in the included readme files in the bin directory, and there are sample DOS batch scripts in the example_batch_scripts directory.

What this is for, and where to look next

LAStools is a specialist toolchain for a specialist data type, and the shape of the project follows from that. LiDAR files are large, they are point based rather than raster based, and the operations that matter are filtering, classifying and rasterising, which is exactly the list of the closed source tools. The open tools are the ones that are about the file format itself, and the licensed ones are about the point processing.

That split is also a licensing decision, and it is worth being clear about it rather than pretending the open tools are the whole product. LASlib and LASzip being open source is what lets a third party build on them, and it is what makes the LAZ format and the ASPRS LAS handling in this repository usable as a foundation. The suite above them is what the company sells.

For a reader trying to place this in the wider LiDAR ecosystem, the formats are the useful part. LAS 1.0 to 1.4 are the ASPRS standard formats, LAZ is the lossless compressed twin of LAS, COPC is the cloud-optimised variant that lascopcindex produces, and the converters cover Terrasolid BIN, E57 for terrestrial scans, ESRI Shapefile, ASCII and the raster formats ASC, BIL, TIF and IMG. A workflow that moves between those formats is what these tools are for, and the lastool, an older GUI for multiple LAStools, is noted as superseded since each tool now has its own GUI.

The documentation is inside the repository rather than on a separate site, in the bin directory readme files, and the Linux specifics are at rapidlasso.de/lastools-linux. The CHANGES.txt file at the root is the changelog for the binaries, and license.html and LICENSE.txt hold the terms, which you should read before deciding whether the open and licensed split affects what you are allowed to redistribute.

Editorial conclusion

LAStools suits a LiDAR practitioner who needs the open tooling for working with point clouds day to day, since laszip, las2las, las2txt, txt2las, lasmerge, lasinfo, lasvalidate and lasindex cover compression, conversion, merging and validation without a licence. It is the wrong choice if you need the rasterising and classification tools, because las2dem, lasground, lasclassify, las2shp and the rest are closed source and licensed, and the download at rapidlasso.de includes all of them rather than only the free ones. Before installing, check the point limit, since the licensed tools can be tested for free up to roughly 3 to 5 million points, and read the Linux dependency list in the README, because the library path has to include the bundled lib directory for the binaries to start.

Frequently asked questions

Official sources

1. Issues
2. LAStools/LAStools on GitHub
3. README
4. Releases