MeshInspector
WebGL required for the interactive viewport
Scan-to-production geometry engine

Scan to Mesh.
Mesh to Production.

One part, one pipeline — from a raw point cloud to a watertight, measured, production-ready model. Scroll to follow it through.

Start free — no install for Web Book a demo
01 — Capture · point cloud

Millions of points, straight from the scanner.

LiDAR, laser, structured-light and photogrammetry data — imported, cleaned and aligned, however it was captured.

02 — Reconstruct · triangle mesh

Rebuilt into a watertight mesh.

Surface reconstruction turns the cloud into clean triangles — tens of millions of them, handled locally.

03 — Repair · solid model

Healed, decimated, production-ready.

Mesh Healer closes holes, self-intersections and thin walls, then simplifies without losing the form.

04 — Verify · deviation & QA

Measured against nominal. Verified.

Tolerance controls, colour deviation maps and an exportable QA report confirm the part is in spec.

Deviation from nominal
−0.100+0.10 mm
1Capture
2Reconstruct
3Repair
4Verify
Scroll ↓
50M+triangles per mesh, processed locally
100M+points per cloud, with CBCT volumes
15+formats — STEP · STL · OBJ · PLY · E57 · LAZ
±10µmverified tolerance control & QA reports
Why teams rely on it

Not a viewer with repair bolted on — a QA-grade pipeline.

01

Verified results

Tolerance controls, colour deviation maps and exportable QA reports confirm a model meets spec before it reaches the machine.
tolerance ±0.01 mm · signed deviation map · PDF report
02

Performance at scale

Handles 10–50M+ triangle meshes, 50–100M+ point clouds and large CBCT datasets without choking — multithreaded and memory-aware.
meshes 50M△ · clouds 100M pts · CBCT volumes
03

Automated & agentic

In-app Python scripting and a built-in MCP server let you batch whole folders or hand the workflow to an AI coding agent.
Python API · MCP server · headless batch
04

Cross-platform

Native desktop apps on Windows, macOS and Linux plus a full WebAssembly build — same engine, same files, everywhere.
native desktop · WebAssembly web build
05

Seamless I/O

Import STEP and convert STL, OBJ, E57, PLY and LAZ with correct unit and axis handling, so geometry round-trips cleanly.
STEP in · STL/OBJ/PLY/E57/LAZ · units & axes
06

Fair pricing

Start free on the web with no install, then scale to desktop and the embeddable SDK when you need automation and throughput.
free web tier · desktop · embeddable SDK
Workflows tailored to your needs

One pipeline, tuned to how you work.

01

Upload CAD file

STEP files convert to meshes locally — nothing leaves your machine.

02

Prepare for production

One-click repair, hole filling and deviation maps against nominal.

03

Engrave text

Emboss serial numbers or identifiers into the geometry.

04

Measure & inspect

Dimensions, GD&T-style checks and a shareable QA report.

05

G-code preview

Verify toolpaths before the part reaches the machine.

repair_and_inspect.py
# Drive MeshInspector from Python or any MCP agent
import meshlib.mrmeshpy as mr

mesh    = mr.loadMesh("scan.ply")
nominal = mr.loadMesh("part_cad.stl")

mr.fixSelfIntersections(mesh)       # watertight
mr.fillHoles(mesh)

report  = mr.deviation(mesh, nominal)   # QA map
print(f"max |dev| = {report.max:.3f} mm")

mr.saveMesh(mesh, "part_ready.stl")
Automated & agentic

Script it, or let an agent run it.

In-app Python — automate repair, decimation, measurement and conversion across whole folders.
MCP server — expose every operation to an AI assistant over the Model Context Protocol.
Headless & embeddable — the same engine ships as the MeshLib SDK for your own pipeline.
Claude CodeClaude DesktopCursorGitHub CopilotWindsurfClineOpenAI CodexGemini CLI
Trusted in the field

What practitioners say.

“It fixed and decimated lattice meshes that every other tool we tried simply gave up on — fast enough to keep in our daily loop.”

Gabriele ImpellizzeriR&D Manager

“Decimation quality is excellent, and the Python API let us wire MeshInspector straight into our production system.”

Peter SekulaChief of Operations, Matterfield

“We automated a workflow that used to be manual by calling the MeshLib API — the throughput gain was immediate and reliable.”

Ruedger RubbertCTO, Brius Technologies
FAQ

Answers before you start.

Which platforms are supported?

MeshInspector runs in the browser and as native desktop apps for Windows, macOS and Linux — same engine and file compatibility across all of them.

Can I use it on mobile?

The web build opens in a mobile browser, but a desktop is recommended for heavy meshes, point clouds and precision measurement.

What data types and formats work?

Meshes, point clouds, volumetric data, polylines, distance maps and G-code — importing STEP and converting STL, OBJ, PLY, E57 and LAZ with correct units and axes.

Is my data secure?

On the web, files stay in your browser's memory; on desktop they never leave your machine. Nothing is uploaded unless you explicitly choose to.

Is there a free trial?

Yes — start free on the web with no install, then move to desktop or the SDK when you need automation and enterprise throughput.

Is it suitable for beginners?

One-click repair, guided workflows and sensible defaults get newcomers to a printable result quickly, while the Python API stays open for experts.

Get started

From scan to production-ready
in one session.

No install for the web build. Bring a scan or a STEP file and export a watertight, measured part.

Start freeBook a demo