NeuroCAD vs nTop: Implicit CAD Showdown

Both NeuroCAD and nTopology (nTop) are implicit CAD platforms. Both represent geometry as fields rather than boundary meshes. Both target advanced manufacturing workflows --- lattice structures, topology optimization, conformal cooling, multi-material design. But the architectures diverge significantly, and those architectural differences determine what each tool is good at, what it struggles with, and who it is for.

This is an honest comparison. We built NeuroCAD, so we are biased. We will state our perspective transparently and let you evaluate.

Architecture

nTop

nTop is a desktop application for Windows built on a proprietary implicit kernel written in C++. The user interface is a node-based visual programming environment (called the “block editor”) where geometry operations are wired together as a dataflow graph. nTop was founded in 2015, has a mature product, and a significant user base in aerospace, medical devices, and motorsport.

The kernel evaluates fields on a voxel grid. Boolean operations, lattice generation, and surface offset all happen as field operations. Meshing (marching cubes or dual contouring) converts the implicit representation to a triangle mesh for export.

NeuroCAD

NeuroCAD is a browser-native CAD platform built on a Rust kernel that compiles to WebAssembly. The kernel uses signed distance fields as the primary representation, augmented with a distributed constraint graph (DCG) for dependency tracking and e-graphs for expression optimization.

The architecture is a 13-crate Rust workspace with strict layered dependencies. The kernel runs in the browser via WASM for interactive use and on the server for heavy computation (batch meshing, simulation). There is no desktop installer.

Key Architectural Differences

Aspect	nTop	NeuroCAD
Language	C++	Rust
Deployment	Desktop (Windows)	Browser (WASM) + server
UI paradigm	Node-based visual programming	Sketch-based + parametric tree
Field representation	Implicit fields on voxel grid	SDF with algebraic composition
Optimization	Manual graph tuning	Automatic e-graph equality saturation
Dependency tracking	Reactive dataflow	DCG with dirty propagation

Sketching and Traditional CAD

nTop

nTop’s strength is in field operations: lattices, offsetting, blending, topology optimization post-processing. Its weakness is traditional sketching. Creating a bracket with precise dimensions, geometric constraints (parallel, perpendicular, tangent, concentric), and a parametric dimension table is awkward in the node editor. nTop is designed to receive geometry from traditional CAD (imported as STEP or STL) and then apply implicit operations to it.

NeuroCAD

NeuroCAD includes a full parametric sketch-to-feature workflow. 2D sketches with geometric constraints and dimensional constraints are solved by a deterministic constraint solver, then extruded, revolved, swept, or lofted into 3D features. This means the engineer can design a part from scratch without importing anything, using the same sketch-dimension-feature workflow they know from SolidWorks or Fusion 360.

The implicit representation kicks in at the feature level. Each feature is an SDF, and features compose through field operations. The result is a hybrid: traditional parametric modeling for human-authored design intent, implicit geometry for lattices, grading, and manufacturing-aware operations.

Our take: If your workflow is “import geometry from Siemens NX, add lattices, export for printing,” nTop is purpose-built for that. If you want a single tool that handles both the initial design and the advanced implicit operations, NeuroCAD covers more ground.

Lattice Design

Both platforms excel at lattice generation, but the approaches differ.

nTop

nTop offers an extensive library of unit cells (gyroid, diamond, Schwarz-P, octet truss, and many more) with GUI controls for cell size, wall thickness, and volume fraction. Lattice conforming (adapting the lattice to fill an arbitrary volume) is mature and reliable. Graded lattices are supported through field-based control of cell parameters.

nTop also provides lattice optimization workflows that connect to FEA results: the local density of the lattice is driven by stress or compliance data from simulation.

NeuroCAD

NeuroCAD represents lattices as mathematical functions evaluated in the SDF framework. TPMS (triply periodic minimal surface) lattices --- gyroid, Schwarz-P, diamond, Lidinoid --- are first-class field primitives. Grading is implemented by modulating lattice parameters with spatial fields, which can themselves be derived from simulation data, distance fields, or user-defined functions.

The e-graph optimizer can simplify lattice expressions. For example, if a lattice is intersected with a half-space and then unioned with a solid block, the optimizer may recognize that the lattice term is zero outside the half-space and eliminate unnecessary evaluations.

Our take: nTop has more unit cell types in the UI and a more polished lattice workflow today. NeuroCAD’s mathematical approach is more flexible for custom unit cells and novel lattice topologies, but requires more technical depth from the user.

Material Fields and Multi-Physics

nTop

nTop supports field-driven design where scalar fields control geometric parameters (thickness, porosity, offset distance). These fields can be imported from simulation (thermal, structural) or defined analytically. nTop’s “field-driven design” is one of its signature capabilities.

NeuroCAD

NeuroCAD’s kernel is built around fields from the ground up. Every geometric entity is a field. Material properties can be represented as additional fields that share the same spatial domain. A density field, a Young’s modulus field, and a thermal conductivity field can all coexist with the geometry field and reference each other through the DCG.

This enables multi-physics-aware design: a thermal simulation produces a temperature field, which drives a porosity field, which controls a lattice, which feeds back into the thermal simulation. The DCG tracks all these dependencies and knows which fields need re-evaluation when a parameter changes.

Our take: Both platforms support field-driven design. NeuroCAD’s DCG provides more explicit dependency tracking, which matters for iterative multi-physics workflows. nTop’s field-driven design is more established and better documented today.

Performance and Scalability

nTop

nTop evaluates fields on the GPU using CUDA. This provides excellent throughput for large models. Meshing and export of complex lattice structures can take minutes to hours depending on resolution, which is typical for implicit CAD.

NeuroCAD

NeuroCAD evaluates fields on the CPU (with SIMD optimization) for the native target and in the browser via WASM. WebGPU compute shaders are being integrated for GPU-accelerated evaluation in the browser. The e-graph optimization pass reduces evaluation depth, which partially compensates for the lack of CUDA-level throughput.

Our take: For raw field evaluation throughput on desktop hardware, nTop’s CUDA backend is faster today. NeuroCAD’s browser deployment removes the need for high-end local hardware --- heavy computation can be offloaded to the server. The performance trade-off depends on your deployment model.

Deployment and Pricing

nTop

nTop is a commercial desktop application with per-seat licensing. Pricing is not publicly listed but is in the range typical of professional engineering software ($15,000—$25,000 per seat per year based on industry reports). Windows only.

NeuroCAD

NeuroCAD runs in the browser. No installation, no OS restriction, no GPU requirement for basic use. The pricing model targets accessibility: tiered plans from free (limited features) to enterprise (full kernel, server compute, API access).

Our take: If your organization already has nTop licenses and a Windows-centric CAD infrastructure, the switching cost is high. If you are starting fresh, need cross-platform access, or want to embed CAD capabilities in a web application, NeuroCAD’s deployment model is more flexible.

Who Should Use What

Choose nTop if:

• Your workflow is import-from-CAD, add-lattice, export-to-print.
• You need the largest library of pre-built unit cells and lattice operations.
• You are already invested in a Windows desktop CAD workflow.
• You need CUDA-accelerated field evaluation for very large models.

Choose NeuroCAD if:

• You want sketch-to-feature modeling and implicit operations in one tool.
• You need browser-native deployment with no installation.
• You want automatic expression optimization via e-graphs.
• You need explicit dependency tracking for multi-physics workflows.
• You prefer open, tiered pricing over enterprise-only licensing.

Final Thoughts

nTop and NeuroCAD are both pushing implicit CAD forward, but from different starting points. nTop started as a lattice and field-driven design tool and is expanding toward general CAD. NeuroCAD started as a full CAD kernel and is building implicit capabilities from the ground up. The convergence point is a tool that handles everything from a 2D sketch to a graded lattice to a manufacturing-ready export --- and both platforms are moving in that direction.

Competition in implicit CAD benefits everyone. The more tools that support field-based geometry, the faster additive manufacturing, medical devices, and advanced materials engineering will advance. We welcome the comparison and look forward to earning engineers’ trust through transparent technical merit.