CLI Anything Cloudcompare

cli-anything-cloudcompare

Agent-friendly command-line harness for CloudCompare — the open-source 3D point cloud and mesh processing software.

41 commands across 9 groups.

Installation

pip install cli-anything-cloudcompare

Prerequisites:

• Python 3.10+
• CloudCompare installed on your system
• Linux (Flatpak): flatpak install flathub org.cloudcompare.CloudCompare
• macOS/Windows: download from https://cloudcompare.org

Tested with: CloudCompare 2.13.2 (Flatpak, Linux)

Global Options

These options must be placed before the subcommand:

cli-anything-cloudcompare [--project FILE] [--json] COMMAND [ARGS]...

Command Groups

1. project — Project Management (3 commands)

project new

Create a new empty project file.

# Create a project with default name
cli-anything-cloudcompare project new -o myproject.json

# Create a project with a custom name
cli-anything-cloudcompare project new -o myproject.json -n "Bridge Survey 2024"

# JSON output for agents
cli-anything-cloudcompare --json project new -o myproject.json

Options: -o/--output TEXT (required), -n/--name TEXT

project info

Show project info and loaded entities.

cli-anything-cloudcompare --project myproject.json project info

# JSON output
cli-anything-cloudcompare --project myproject.json --json project info

project status

Show quick project status (cloud count, mesh count, last operation).

cli-anything-cloudcompare --project myproject.json project status

2. session — Session Management (4 commands)

session save

Save the current project state to disk.

cli-anything-cloudcompare --project myproject.json session save

session history

Show recent operation history.

# Show last 10 operations (default)
cli-anything-cloudcompare --project myproject.json session history

# Show last 5 operations
cli-anything-cloudcompare --project myproject.json session history -n 5

Options: -n/--last INTEGER

session set-format

Update the default export format for future operations.

# Set default cloud export to LAS
cli-anything-cloudcompare --project myproject.json session set-format --cloud-fmt LAS --cloud-ext las

# Set default mesh export to OBJ
cli-anything-cloudcompare --project myproject.json session set-format --mesh-fmt OBJ --mesh-ext obj

# Set both cloud and mesh defaults
cli-anything-cloudcompare --project myproject.json session set-format \
  --cloud-fmt PLY --cloud-ext ply \
  --mesh-fmt STL --mesh-ext stl

Options: --cloud-fmt TEXT, --cloud-ext TEXT, --mesh-fmt TEXT, --mesh-ext TEXT

session undo

Remove the last operation from history (soft undo — does not delete output files).

cli-anything-cloudcompare --project myproject.json session undo

3. cloud — Point Cloud Operations (21 commands)

All cloud commands take CLOUD_INDEX (0-based integer from cloud list) and most accept --add-to-project to register the output back into the project.

cloud add

Add a point cloud file to the project.

# Add a LAS file
cli-anything-cloudcompare --project myproject.json cloud add /data/scan.las

# Add with a label
cli-anything-cloudcompare --project myproject.json cloud add /data/scan.las -l "roof scan"

Options: -l/--label TEXT

cloud list

List all clouds currently in the project.

cli-anything-cloudcompare --project myproject.json cloud list

# JSON output for parsing indices
cli-anything-cloudcompare --project myproject.json --json cloud list

cloud convert

Convert a cloud from one format to another (format determined by file extension).

# LAS → PLY
cli-anything-cloudcompare cloud convert /data/scan.las /data/scan.ply

# PCD → LAS
cli-anything-cloudcompare cloud convert /data/cloud.pcd /data/cloud.las

cloud subsample

Reduce the number of points using RANDOM, SPATIAL, or OCTREE method.

# Random: keep 100 000 points
cli-anything-cloudcompare --project myproject.json cloud subsample 0 \
  -o /data/sub_random.las -m random -n 100000

# Spatial: minimum distance 0.05 m between points
cli-anything-cloudcompare --project myproject.json cloud subsample 0 \
  -o /data/sub_spatial.las -m spatial -n 0.05

# Octree: level 8
cli-anything-cloudcompare --project myproject.json cloud subsample 0 \
  -o /data/sub_octree.las -m octree -n 8 --add-to-project

Options: -o/--output TEXT (required), -m/--method [random|spatial|octree], -n/--param FLOAT, --add-to-project

cloud crop

Crop a cloud to an axis-aligned bounding box.

# Keep points inside the box
cli-anything-cloudcompare --project myproject.json cloud crop 0 \
  -o /data/cropped.las \
  --xmin 0.0 --ymin 0.0 --zmin 0.0 \
  --xmax 10.0 --ymax 10.0 --zmax 5.0

# Keep points OUTSIDE the box
cli-anything-cloudcompare --project myproject.json cloud crop 0 \
  -o /data/exterior.las \
  --xmin 0.0 --ymin 0.0 --zmin 0.0 \
  --xmax 10.0 --ymax 10.0 --zmax 5.0 --outside

Options: -o/--output TEXT (required), --xmin/ymin/zmin/xmax/ymax/zmax FLOAT (all required), --outside, --add-to-project

cloud normals

Compute surface normals via the octree method.

# Compute normals at octree level 6
cli-anything-cloudcompare --project myproject.json cloud normals 0 \
  -o /data/with_normals.ply --level 6

# Compute normals oriented toward +Z
cli-anything-cloudcompare --project myproject.json cloud normals 0 \
  -o /data/with_normals.ply --level 6 --orientation plus_z --add-to-project

Options: -o/--output TEXT (required), --level INTEGER (1–10), --orientation [plus_x|plus_y|plus_z|minus_x|minus_y|minus_z], --add-to-project

cloud invert-normals

Flip all normal vectors in the cloud.

cli-anything-cloudcompare --project myproject.json cloud invert-normals 0 \
  -o /data/flipped_normals.ply --add-to-project

Options: -o/--output TEXT (required), --add-to-project

cloud filter-sor

Statistical Outlier Removal — removes isolated noise points.

# Default parameters (k=6 neighbours, 1.0 std ratio)
cli-anything-cloudcompare --project myproject.json cloud filter-sor 0 \
  -o /data/denoised.las

# Custom parameters
cli-anything-cloudcompare --project myproject.json cloud filter-sor 0 \
  -o /data/denoised.las --nb-points 12 --std-ratio 2.0 --add-to-project

Options: -o/--output TEXT (required), --nb-points INTEGER, --std-ratio FLOAT, --add-to-project

cloud noise-filter

Remove noisy points using the PCL noise filter (KNN or radius mode).

# KNN mode (default)
cli-anything-cloudcompare --project myproject.json cloud noise-filter 0 \
  -o /data/clean.las --knn 8 --noisiness 1.0

# Radius mode
cli-anything-cloudcompare --project myproject.json cloud noise-filter 0 \
  -o /data/clean.las --radius 0.1 --use-radius --add-to-project

Options: -o/--output TEXT (required), --knn INTEGER, --noisiness FLOAT, --radius FLOAT, --use-radius, --absolute, --add-to-project

cloud filter-csf

Ground filtering using the Cloth Simulation Filter (CSF) algorithm. Separates ground from off-ground points (buildings, vegetation).

# Extract ground only — mixed terrain
cli-anything-cloudcompare --project myproject.json cloud filter-csf 0 \
  --ground /data/ground.las --scene relief

# Split ground + off-ground — urban scene
cli-anything-cloudcompare --project myproject.json cloud filter-csf 0 \
  --ground /data/ground.las \
  --offground /data/buildings.las \
  --scene flat --cloth-resolution 0.5 --class-threshold 0.3

# Steep forested slope with slope post-processing
cli-anything-cloudcompare --project myproject.json cloud filter-csf 0 \
  --ground /data/terrain.las --scene slope --proc-slope --add-to-project