Skip to content

For Companies

invertmeeg is a Python library for M/EEG inverse solutions with a unified API, a large solver catalog, and a built-in benchmarking stack. It integrates with the mne-python ecosystem and returns standard mne.SourceEstimate objects.

This page is for teams who want to use or ship inverse methods in commercial software without taking on the obligations of GPLv3.

Typical users

  • Neurotech & neurofeedback teams working with low-channel EEG and tight runtime constraints
  • M/EEG software vendors who want a broad set of inverse methods behind one stable interface
  • R&D groups evaluating methods, building new ones, or standardizing internal pipelines

What you get

  • A single entry point (Solver("solver_id")) across many solver families (minimum norm, beamformers, Bayesian, sparse recovery, subspace, optional deep learning).
  • Simulation, evaluation, and benchmarking utilities to compare methods under controlled conditions.
  • Repeatable artifacts (JSON benchmark outputs) that can be turned into dashboards, reports, and internal decision docs.

Why teams adopt invertmeeg

  • Faster integration: one API and consistent outputs instead of many disconnected implementations.
  • Method selection you can justify: benchmark solvers on scenarios that match your product (channels, SNR, focal vs. extended sources).
  • Room for productization: the core stays open for science, while commercial licensing supports closed-source distribution.

Licensing (commercial use)

invertmeeg is dual-licensed:

  • GPLv3 (open source): for research, education, and open-source distribution under GPLv3.
  • Commercial license (closed source): for proprietary/closed-source products or internal distribution that can’t meet GPLv3 obligations.

To discuss a commercial license, email: lukas.hecker.job@gmail.com.

Intended use / compliance

invertmeeg is provided as a research and engineering tool. If you are building software in a regulated medical context, you are responsible for validating performance, managing risk, and meeting any applicable regulatory and quality requirements.

Next step

If you share (1) your channel setup, (2) a representative forward model (or your constraints), and (3) your target use case, we can define a small, repeatable benchmark that answers: “which solver family is a good default for our product?”