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
bash
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:
bash
cli-anything-cloudcompare [--project FILE] [--json] COMMAND [ARGS]...
| Option | Description |
|---|
| Path to project JSON file |
| Output results as JSON (for agent consumption) |
Command Groups
1. project — Project Management (3 commands)
project new
Create a new empty project file.
bash
# 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
project info
Show project info and loaded entities.
bash
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).
bash
cli-anything-cloudcompare --project myproject.json project status
2. session — Session Management (4 commands)
session save
Save the current project state to disk.
bash
cli-anything-cloudcompare --project myproject.json session save
session history
Show recent operation history.
bash
# 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
session set-format
Update the default export format for future operations.
bash
# 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
session undo
Remove the last operation from history (soft undo — does not delete output files).
bash
cli-anything-cloudcompare --project myproject.json session undo
3. cloud — Point Cloud Operations (21 commands)
All cloud commands take
(0-based integer from
) and most accept
to register the output back into the project.
cloud add
Add a point cloud file to the project.
bash
# 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"
cloud list
List all clouds currently in the project.
bash
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).
bash
# 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.
bash
# 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:
(required),
-m/--method [random|spatial|octree]
,
,
cloud crop
Crop a cloud to an axis-aligned bounding box.
bash
# 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:
(required),
--xmin/ymin/zmin/xmax/ymax/zmax FLOAT
(all required),
,
cloud normals
Compute surface normals via the octree method.
bash
# 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:
(required),
(1–10),
--orientation [plus_x|plus_y|plus_z|minus_x|minus_y|minus_z]
,
cloud invert-normals
Flip all normal vectors in the cloud.
bash
cli-anything-cloudcompare --project myproject.json cloud invert-normals 0 \
-o /data/flipped_normals.ply --add-to-project
cloud filter-sor
Statistical Outlier Removal — removes isolated noise points.
bash
# 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
cloud noise-filter
Remove noisy points using the PCL noise filter (KNN or radius mode).
bash
# 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:
(required),
,
,
,
,
,
cloud filter-csf
Ground filtering using the Cloth Simulation Filter (CSF) algorithm. Separates ground from off-ground points (buildings, vegetation).
bash
# 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
Options:
(required),
,
--scene [slope|relief|flat]
,
,
,
,
,
cloud filter-sf
Filter a cloud by scalar field value range (keep points where SF ∈ [min, max]).
bash
# Keep points with SF value between 10 and 50
cli-anything-cloudcompare --project myproject.json cloud filter-sf 0 \
-o /data/filtered.las --min 10.0 --max 50.0
# Filter using a specific SF index
cli-anything-cloudcompare --project myproject.json cloud filter-sf 0 \
-o /data/filtered.las --min 0.0 --max 1.5 --sf-index 2 --add-to-project
Options:
(required),
(required),
(required),
,
cloud sf-from-coord
Convert a coordinate axis (X/Y/Z) to a scalar field. Commonly used to create a height (Z) scalar field.
bash
# Create Z scalar field (height)
cli-anything-cloudcompare --project myproject.json cloud sf-from-coord 0 \
-o /data/with_z_sf.las --dim z --add-to-project
# Create X scalar field with a specific active index
cli-anything-cloudcompare --project myproject.json cloud sf-from-coord 0 \
-o /data/with_x_sf.las --dim x --sf-index 0
Options:
(required),
(default: z),
,
cloud sf-filter-z
Convenience command: convert Z → scalar field and filter by height range in one step.
bash
# Extract points between z=1.0 m and z=2.5 m
cli-anything-cloudcompare --project myproject.json cloud sf-filter-z 0 \
-o /data/slice.las --min 1.0 --max 2.5 --add-to-project
# Only apply upper bound
cli-anything-cloudcompare --project myproject.json cloud sf-filter-z 0 \
-o /data/below_5m.las --max 5.0
cloud sf-to-rgb
Convert the active scalar field to RGB colours.
bash
cli-anything-cloudcompare --project myproject.json cloud sf-to-rgb 0 \
-o /data/coloured.ply --add-to-project
cloud rgb-to-sf
Convert RGB colours to a scalar field (luminance value).
bash
cli-anything-cloudcompare --project myproject.json cloud rgb-to-sf 0 \
-o /data/luminance.las --add-to-project
cloud curvature
Compute curvature scalar field (MEAN or GAUSS).
bash
# Mean curvature with radius 0.5 m
cli-anything-cloudcompare --project myproject.json cloud curvature 0 \
-o /data/curvature.las --type mean --radius 0.5
# Gaussian curvature
cli-anything-cloudcompare --project myproject.json cloud curvature 0 \
-o /data/curvature.las --type gauss --radius 0.5 --add-to-project
cloud roughness
Compute roughness scalar field (deviation from local best-fit plane).
bash
cli-anything-cloudcompare --project myproject.json cloud roughness 0 \
-o /data/roughness.las --radius 0.2 --add-to-project
cloud density
Compute point density scalar field.
bash
# KNN density
cli-anything-cloudcompare --project myproject.json cloud density 0 \
-o /data/density.las --type knn --radius 0.5
# Surface density
cli-anything-cloudcompare --project myproject.json cloud density 0 \
-o /data/density.las --type surface --radius 1.0 --add-to-project
Options:
(required),
,
--type [knn|surface|volume]
,
cloud segment-cc
Segment cloud into connected components (clusters). Each component is saved as a separate file.
bash
# Segment with octree level 8, minimum 100 points per component
cli-anything-cloudcompare --project myproject.json cloud segment-cc 0 \
-o /data/components/ --octree-level 8 --min-points 100
# Save components as PLY files
cli-anything-cloudcompare --project myproject.json cloud segment-cc 0 \
-o /data/components/ --octree-level 6 --min-points 50 --fmt ply
cloud merge
Merge all clouds in the project into a single cloud.
bash
cli-anything-cloudcompare --project myproject.json cloud merge \
-o /data/merged.las --add-to-project
cloud mesh-delaunay
Build a 2.5-D Delaunay triangulation mesh from a cloud.
bash
# Basic Delaunay mesh
cli-anything-cloudcompare --project myproject.json cloud mesh-delaunay 0 \
-o /data/surface.obj
# Best-fit plane with max edge length limit
cli-anything-cloudcompare --project myproject.json cloud mesh-delaunay 0 \
-o /data/surface.ply --best-fit --max-edge-length 2.0 --add-to-project
4. mesh — Mesh Operations (3 commands)
mesh add
Add a mesh file to the project.
bash
cli-anything-cloudcompare --project myproject.json mesh add /data/model.obj
# Add with label
cli-anything-cloudcompare --project myproject.json mesh add /data/model.ply -l "building model"
mesh list
List all meshes in the project.
bash
cli-anything-cloudcompare --project myproject.json mesh list
# JSON output
cli-anything-cloudcompare --project myproject.json --json mesh list
mesh sample
Sample a point cloud from a mesh surface.
bash
# Sample 50 000 points from mesh at index 0
cli-anything-cloudcompare --project myproject.json mesh sample 0 \
-o /data/sampled.las -n 50000
# Add sampled cloud back to project
cli-anything-cloudcompare --project myproject.json mesh sample 0 \
-o /data/sampled.las -n 100000 --add-to-project
5. distance — Distance Computation (2 commands)
distance c2c
Compute cloud-to-cloud distances. Adds a distance scalar field to the compared cloud.
bash
# Compare cloud 1 to reference cloud 0
cli-anything-cloudcompare --project myproject.json distance c2c \
--compare 1 --reference 0 -o /data/distances.las
# Split into X/Y/Z components at octree level 8
cli-anything-cloudcompare --project myproject.json distance c2c \
--compare 1 --reference 0 -o /data/distances.las \
--split-xyz --octree-level 8 --add-to-project
Options:
(required),
(required),
(required),
,
,
distance c2m
Compute cloud-to-mesh distances. Adds a distance scalar field to the cloud.
bash
# Basic cloud-to-mesh distance
cli-anything-cloudcompare --project myproject.json distance c2m \
--cloud 0 --mesh 0 -o /data/c2m_dist.las
# With flipped normals and unsigned distances
cli-anything-cloudcompare --project myproject.json distance c2m \
--cloud 0 --mesh 0 -o /data/c2m_dist.las \
--flip-normals --unsigned --add-to-project
Options:
(required),
(required),
(required),
,
,
6. transform — Transformations and Registration (2 commands)
transform apply
Apply a 4×4 rigid-body transformation matrix to a cloud.
bash
# Apply a transformation matrix from file
cli-anything-cloudcompare --project myproject.json transform apply 0 \
-o /data/transformed.las -m /data/matrix.txt
# Apply the inverse transformation
cli-anything-cloudcompare --project myproject.json transform apply 0 \
-o /data/transformed.las -m /data/matrix.txt --inverse --add-to-project
The matrix file must contain 4 rows of 4 space-separated values:
1 0 0 0
0 1 0 0
0 0 1 0
0 0 0 1
Options:
(required),
(required),
,
transform icp
Run ICP (Iterative Closest Point) registration to align one cloud to another.
bash
# Basic ICP alignment
cli-anything-cloudcompare --project myproject.json transform icp \
--aligned 1 --reference 0 -o /data/aligned.las
# ICP with overlap and iteration control
cli-anything-cloudcompare --project myproject.json transform icp \
--aligned 1 --reference 0 -o /data/aligned.las \
--max-iter 50 --overlap 80 --min-error-diff 1e-6 --add-to-project
Options:
(required),
(required),
(required),
,
,
,
7. export — Export Clouds and Meshes (4 commands)
export formats
List all available export format presets.
bash
cli-anything-cloudcompare export formats
# JSON output
cli-anything-cloudcompare --json export formats
export cloud
Export a cloud to a target format.
bash
# Export cloud at index 0 as LAS
cli-anything-cloudcompare --project myproject.json export cloud 0 /data/output.las
# Export as PLY using preset
cli-anything-cloudcompare --project myproject.json export cloud 0 /data/output.ply -f ply
# Overwrite if file exists
cli-anything-cloudcompare --project myproject.json export cloud 0 /data/output.las -f las --overwrite
Supported presets:
,
,
,
,
,
,
,
,
export mesh
Export a mesh to a target format.
bash
# Export mesh at index 0 as OBJ
cli-anything-cloudcompare --project myproject.json export mesh 0 /data/model.obj
# Export as STL
cli-anything-cloudcompare --project myproject.json export mesh 0 /data/model.stl -f stl --overwrite
export batch
Batch export all project clouds to a directory.
bash
# Export all clouds as LAS
cli-anything-cloudcompare --project myproject.json export batch \
-d /data/exports/ -f las
# Overwrite existing files
cli-anything-cloudcompare --project myproject.json export batch \
-d /data/exports/ -f ply --overwrite
8. info — Installation Info (1 command)
Show CloudCompare installation path and version.
bash
cli-anything-cloudcompare info
# JSON output
cli-anything-cloudcompare --json info
9. repl — Interactive REPL (1 command)
Start the interactive REPL session with history and undo support.
bash
# Start REPL without a project
cli-anything-cloudcompare repl
# Start REPL with an existing project
cli-anything-cloudcompare repl -p myproject.json
# Equivalent: run without subcommand
cli-anything-cloudcompare --project myproject.json
Inside the REPL, type
to list available commands or
to revert the last operation.
Supported File Formats
| Format | Extension | Read | Write | Notes |
|---|
| LAS | | ✓ | ✓ | LiDAR standard, supports intensity/RGB |
| LAZ | | ✓ | ✓ | Compressed LAS |
| PLY | | ✓ | ✓ | ASCII or binary |
| PCD | | ✓ | ✓ | PCL format |
| XYZ | | ✓ | ✓ | Plain text XYZ |
| ASC | | ✓ | ✓ | ASCII with header |
| CSV | | ✓ | ✓ | Comma-separated |
| E57 | | ✓ | ✓ | ASTM scanner exchange |
| BIN | | ✓ | ✓ | CloudCompare native binary |
| OBJ | | ✓ | ✓ | Mesh (Wavefront) |
| STL | | ✓ | ✓ | Mesh (3D printing) |
Typical Workflows
Workflow 1: LiDAR Pre-processing Pipeline
bash
P=myproject.json
# 1. Create project and load scan
cli-anything-cloudcompare project new -o $P
cli-anything-cloudcompare --project $P cloud add /data/scan.las
cli-anything-cloudcompare --project $P cloud list # note index → 0
# 2. Denoise
cli-anything-cloudcompare --project $P cloud filter-sor 0 \
-o /data/denoised.las --nb-points 6 --std-ratio 1.0 --add-to-project
# 3. Subsample to 5 cm grid
cli-anything-cloudcompare --project $P cloud subsample 1 \
-o /data/subsampled.las -m spatial -n 0.05 --add-to-project
# 4. Extract ground plane (CSF)
cli-anything-cloudcompare --project $P cloud filter-csf 2 \
--ground /data/ground.las --offground /data/objects.las \
--scene relief --add-to-project
# 5. Export result
cli-anything-cloudcompare --project $P export cloud 3 /data/ground_final.las -f las --overwrite
Workflow 2: Change Detection Between Two Scans
bash
P=compare.json
cli-anything-cloudcompare project new -o $P
cli-anything-cloudcompare --project $P cloud add /data/scan_2023.las # index 0
cli-anything-cloudcompare --project $P cloud add /data/scan_2024.las # index 1
# ICP alignment (align 2024 to 2023)
cli-anything-cloudcompare --project $P transform icp \
--aligned 1 --reference 0 -o /data/aligned_2024.las \
--overlap 90 --add-to-project # index 2
# Cloud-to-cloud distance
cli-anything-cloudcompare --project $P distance c2c \
--compare 2 --reference 0 -o /data/change_map.las --add-to-project
# Export as LAS with distance scalar field
cli-anything-cloudcompare --project $P export cloud 3 /data/change_map_final.las --overwrite
Workflow 3: Height Slice Extraction
bash
P=slice.json
cli-anything-cloudcompare project new -o $P
cli-anything-cloudcompare --project $P cloud add /data/building.las
# Extract points at 2–3 m height (floor level)
cli-anything-cloudcompare --project $P cloud sf-filter-z 0 \
-o /data/floor_slice.las --min 2.0 --max 3.0 --add-to-project
# Export
cli-anything-cloudcompare --project $P export cloud 1 /data/floor_slice_out.las --overwrite
Workflow 4: Surface Reconstruction
bash
P=mesh.json
cli-anything-cloudcompare project new -o $P
cli-anything-cloudcompare --project $P cloud add /data/terrain.las
# Compute normals
cli-anything-cloudcompare --project $P cloud normals 0 \
-o /data/with_normals.ply --level 6 --orientation plus_z --add-to-project
# Delaunay mesh
cli-anything-cloudcompare --project $P cloud mesh-delaunay 1 \
-o /data/terrain_mesh.obj --max-edge-length 1.0 --add-to-project
# Export mesh
cli-anything-cloudcompare --project $P export mesh 0 /data/terrain_mesh_final.obj --overwrite
Error Handling
| Exit Code | Meaning |
|---|
| Success |
| General error (see stderr for details) |
| Invalid arguments |
Common errors:
bash
# CloudCompare not found
# → Install CloudCompare; check `cli-anything-cloudcompare info`
# Index out of range
# → Run `cloud list` or `mesh list` to confirm valid indices
# File already exists (no --overwrite)
# → Add --overwrite flag to export commands
# fcntl not available (Windows)
# → File locking is skipped automatically; project save still works
For AI Agents
- Always use flag for parseable output
- Check return codes — 0 for success, non-zero for errors
- Parse stderr for error messages on failure
- Use absolute paths for all file arguments
- Verify output files exist after export operations
- Chain with to build multi-step pipelines without re-loading files
- Use to discover valid cloud indices before each operation
- Use to discover available format presets
Version
| Component | Version |
|---|
| cli-anything-cloudcompare | 1.0.0 |
| CloudCompare (tested) | 2.13.2 |
| Python (minimum) | 3.10 |