Many foundational neuroimaging tools are free for academic use, but employ restrictive licenses that hinder commercial application and deter external developers who worry about intellectual property contamination. We clean-room reimplement these core algorithms and distribute our software under permissive licenses (BSD and MIT). We build freedom software. Our code is free as in speech: you can integrate, modify, and deploy it without licensing friction or legal ambiguity.

Beyond licensing, many domain tools suffer from performance bottlenecks. We address this on three fronts:

Native Execution

Porting high-level scripts to native code delivers order-of-magnitude speedups.

Modern Hardware Optimization

Legacy codebases often miss modern compiler and architecture features. We rewrite performance-critical paths to leverage SIMD vectorization, FMA instructions, OpenMP parallelism, and cache-conscious memory layouts.

Cross-Vendor GPU Acceleration

Rather than locking workflows into proprietary platforms like NVIDIA CUDA, we optimize our graphics and compute pipelines across WebGPU, Metal, and Vulkan to deliver high performance on AMD, Apple, Intel, and NVIDIA hardware alike.

Our software aims for functional equivalence with the standard tools the community relies on, but with a significantly smaller footprint, fewer dependencies, and no restrictive licensing. Note that “equivalent” does not mean bit-identical: modern optimizations (such as fused multiply-add operations) preserve higher precision during intermediate calculations than traditional sequential floating-point operations.