Skip to main content
v2026.11,986 entries · CC-BY 4.0

OMFIT Fusion Modeling Framework: Access, Modules, Use

OMFIT is a Python framework for tokamak modeling and data analysis: how access works, what it excludes and which physics modules it ships.

Written and maintained by CASRAI Editorial Board

Published

On this page

OMFIT is an integrated modeling and experimental data analysis software for magnetically confined thermonuclear fusion experiments, developed by O. Meneghini and collaborators. Its name stands for One Modeling Framework for Integrated Tasks. Access is granted after a users agreement form is completed, and the physics codes that its modules call are installed separately. It is not the fitness program of a similar name that dominates web results for the bare word.

Short answer: what OMFIT is

OMFIT is a Python framework that lets fusion researchers link experimental data retrieval, analysis scripts, equilibrium reconstruction and transport or stability simulations into one reproducible project. The project site describes its goal as enhancing existing scientific workflows and enabling new integrated modeling capabilities, and says it pursues that through a bottom-up collaborative development approach (OMFIT project site, read 2026-11-19). It is a research tool for plasma physics groups, not a consumer product, and it has no connection to gym or wellness programs of the same spelling.

Definition and scope

Magnetic-confinement fusion research produces a chain of dependent steps: pull measurements from an experiment, reconstruct the plasma equilibrium, fit kinetic profiles, run a heating or transport code, then compare the result with data. Each step is often a separate code written by a different group. OMFIT sits above these codes as a workflow and data-analysis layer. The project site reports that the framework is used as the framework of the Advanced TOkamak (AToM) project.

The early papers set out the aim. The 2013 paper “Integrated Modeling of Tokamak Experiments with OMFIT” by Orso Meneghini and Lang Lao appeared in Plasma and Fusion Research (DOI 10.1585/pfr.8.2403009), and “Integrated modeling applications for tokamak experiments with OMFIT” by O. Meneghini, S.P. Smith, L.L. Lao and co-authors appeared in Nuclear Fusion in 2015 (DOI 10.1088/0029-5515/55/8/083008). Both records were read through the Crossref registry on 2026-11-19. The project footer carries the copyright line “2012-2026 O. Meneghini and collaborators”, which is the closest thing to a founder statement the project publishes.

What OMFIT is not: it is not a bundled suite of physics solvers. The install documentation states that the OMFIT source code does not include the programs its modules execute externally, and names EFIT, ONETWO, TRANSP, TGYRO, ELITE, SOLPS, M3D-C1 and GPEC as examples (OMFIT install page, read 2026-11-19). A module that runs a separate code has to be pointed at a personal installation of that code, or tunnelled to a server where it is available.

How OMFIT is used in practice

Getting access

The install page lists the steps for the source code: generate an SSH key, obtain a GitHub account and register the key, obtain a GitLab account and register the key, fill out the OMFIT users agreement form with your GitHub username, and then confirm you can open the repository at github.com/gafusion/OMFIT-source. The documented clone command uses the unstable branch. The project site makes the same point on its front page: fill out the users agreement form to start using OMFIT and obtain access to the repository. Because access runs through an agreement, a research office should treat OMFIT as a managed-access scientific tool rather than an anonymous open download.

Environments

The install page states that Python 2.7 support has been dropped and all new installations need Python 3.6 or later. It recommends the Anaconda package manager on Linux and macOS, the Windows Subsystem for Linux with Anaconda on Windows, and notes that OMFIT has been installed on a Raspberry Pi 4 through pip requirements and that a Docker image runs wherever Docker can be installed. Launch scripts are provided for each route, for example a conda script and separate Docker scripts for Linux, macOS and Windows.

Public installations and first run

Many users never install it themselves. The Run page lists 19 publicly available installations maintained at institutions (table below): 7 in the Americas (for example the DIII-D site at General Atomics, PPPL for NSTX-U, MIT PSFC for C-Mod, NERSC), 4 in Asia (ASIPP for EAST, CFETR, SWIP, KFE for KSTAR) and 8 in Europe (for example CCFE for JET and MAST, MPI-IPP for ASDEX, EPFL for TCV, ITER, WEST). On first start, users accept the users agreement form and enter an email address in the Preferences window (OMFIT Run page, read 2026-11-19).

Where the public installations are

Region Listed installations Examples named on the Run page
Americas 7 GA (DIII-D), PPPL (NSTX-U), MIT-PSFC (C-Mod), NERSC
Asia 4 ASIPP (EAST), CFETR, SWIP (HL2A and HL2M), KFE (KSTAR)
Europe 8 CCFE (JET and MAST), MPI-IPP (ASDEX), EPFL (TCV), ITER, WEST

These counts come from the lists on the project’s Run page as listed on 2026-11-19. They are installations maintained for researchers at those institutions, not a count of users.

As a Python library

The Run page also documents using OMFIT as a Python library after installing it as a module, and the install page states that the OMFIT classes can be installed standalone with pip install --upgrade omfit_classes, for use in IPython or Jupyter sessions without loading the whole framework. This is the route most relevant to people who only need file readers and analysis classes, such as the equilibrium-file class used in the Run page’s own example.

Module catalogue: what the physics modules cover

The Modules page lists 109 numbered modules that are publicly available as part of the OMFIT source repository, and states that the list excludes modules private to users or not released by institutions. Of those 109 entries, 54 carry a tutorial flag on the page. The groupings below follow the module descriptions and keywords on that page; the data table after the Sources section compares twenty of them.

  • Equilibrium and stability: modules that run EFIT, CHEASE, HELENA, ELITE, BALOO, GATO, MARS, M3D-C1, NIMROD and GPEC.
  • Transport and turbulence: modules for TRANSP, ONETWO, GYRO, NEO, TGLF and TGYRO, and a transport solver module.
  • Heating, fuelling and fast ions: FREYA, GENRAY, TORAY, CQL3D, PAM, RABBIT and viewers for NUBEAM output.
  • Diagnostics and data: beam emission spectroscopy, Thomson scattering, electron cyclotron emission imaging, Langmuir probes, charge exchange and magnetics analysis.
  • Boundary and divertor physics: a two-point scrape-off-layer model, SOLPS, DEGAS2, detachment studies and heat-flux analysis.
  • Frameworks and workflow: modules that run the FUSE and IPS integrated-modeling frameworks, IMAS-based workflows, and a TUTORIAL module of sample scripts.

The project site’s own description frames OMFIT around experiments, so the modules mix two kinds of work: running simulation codes and analysing measured data. The tutorial flags on the Modules page show that about half of the public modules (54 of 109) come with a worked tutorial, which helps a new group judge how much onboarding effort a module needs before relying on it. A module without a flagged tutorial is not necessarily unsupported; the install page directs users to the module developers for assistance with codes that run separately.

OMFIT is easiest to understand next to the codes it orchestrates. EFIT reconstructs the equilibrium from magnetic data; TRANSP and ONETWO are transport codes; TGLF and TGYRO model turbulent transport; SOLPS models the plasma edge. OMFIT does not replace them. It supplies the scripted glue that moves their inputs and outputs between steps.

That role connects directly to research-software practice. A workflow that chains many external executables is only reproducible if the versions and inputs are recorded, which is the concern of computational reproducibility and the broader idea of reproducibility. When a group publishes results produced this way, the software itself needs to be cited, a practice described in the FAIR4RS software citation principles. Teams that want their analysis code described in a peer-reviewed venue can see the Journal of Open Research Software guide, and internal quality checks are covered under code review for research software.

Where it matters for labs and research offices

  • Access governance. Entry runs through a users agreement and a GitHub repository permission, so onboarding a new group member requires an account and an approved request, not just an installer.
  • External code dependencies. Because the source does not include codes such as TRANSP or SOLPS, every module that depends on one needs its own installation or remote server, each with its own licensing and contact point.
  • Shared installations. The 19 listed public installations mean many researchers can run OMFIT at a host institution or facility without a personal setup.
  • Documentation of results. The 2013 and 2015 papers are the references the project’s own literature points to when describing the framework, and a methods section should record the OMFIT version, shown on the project site as 3.2026.43 on 2026-11-19.

Common misunderstandings

“OMFIT” and the fitness program

The bare word returns mostly social-media and app pages for an unrelated fitness program, along with questions such as whether OMAD diets pair with workouts or what an Omnia fitness class is. None of that concerns the fusion software on this page. For fusion modeling, search for “OMFIT fusion modeling framework” or the full name One Modeling Framework for Integrated Tasks.

“It is a simulator”

OMFIT is not itself a physics solver. It runs, links and visualizes other codes and analysis scripts, and it ships its own analysis modules for diagnostics and profile fitting.

“It is a public download”

The project site and install page route access through the users agreement form and repository permission, with the unstable branch as the documented clone target. Only the OMFIT classes can be installed standalone through pip.

“The module list shows everything in use”

The Modules page says it lists modules publicly available in the source repository and excludes private and unreleased ones, so it understates what an individual institution may run.

Frequently asked questions

Who is the founder of OMFIT?

For the fusion software, the project site credits “O. Meneghini and collaborators” in its 2012-2026 copyright line, and the 2013 Plasma and Fusion Research paper lists Orso Meneghini and Lang Lao as authors. The project calls its approach bottom-up and collaborative, so it is not a single-author tool.

How do I get access to OMFIT?

Fill out the OMFIT users agreement form, give your GitHub username so you can be added to the repository, then clone the repository as the install page describes. Alternatively, use one of the public installations listed on the Run page.

Does OMFIT include EFIT, TRANSP and the other physics codes?

No. The install page says the source code does not include externally executed programs such as EFIT, ONETWO, TRANSP, TGYRO, ELITE, SOLPS, M3D-C1 and GPEC.

Can I use OMFIT classes in my own Python code?

Yes. The install page documents pip install --upgrade omfit_classes for standalone use of the classes, for example in Jupyter notebooks.

Is OMFIT the same as the OMFIT fitness program, “OMFIT for parents” or the Omnia fitness class?

No. Those are unrelated consumer fitness topics. This page covers only the magnetic-fusion software, and the same applies to questions about OMAD diets.

Sources

Data: Twenty public OMFIT physics modules and whether a tutorial is flagged

Selection of twenty of the 109 publicly listed OMFIT modules from the project's Modules page, giving each module's published description and whether the page flags a tutorial. Compiled from the OMFIT Modules page read on 2026-11-19.

Period covered: 2026.

Twenty public OMFIT physics modules and whether a tutorial is flagged
ModulePublished descriptionTutorial flagged on Modules page
EFITRun the EFIT equilibrium solverYes
TRANSPTokamak whole device model and transport code from PPPLYes
ELITERun the ELITE ideal MHD stability codeYes
SOLPSRuns the SOLPS code, displays output, and compares runsYes
M3DC1Workflow for automated running of M3D-C1Yes
GPECGeneralized Perturbed Equilibrium Code: MHD stability, plasma response, torque from 3D fieldsYes
TGLF_GACODERuns the TGLF code (reduced gyrokinetic turbulence transport model)Yes
TGYRO_GACODERuns the steady state TGYRO transport codeYes
ONETWOtimeRun the ONETWO transport codeYes
NIMRODRun the NIMROD MHD codeYes
BESBeam Emission Spectroscopy analysis toolboxYes
ECEIECEI diagnostic dataYes
OMFITprofilesAdvanced fitting of kinetic profilesYes
STEPStability, Transport, Equilibrium, and Pedestal workflowYes
TUTORIALCollection of scripts illustrating common operations in OMFITYes
2POINTSOL2 point SOL modelNo
BALOORuns the BALOO infinite-n stability codeNo
CHEASERun the CHEASE equilibrium codeNo
FREYARun the FREYA neutral beam codeNo
GENRAYRun the GENRAY ray tracing codeNo

Sources

Compiled by CASRAI from the sources above. Free to reuse under CC BY 4.0 with attribution.

Follow CASRAI

Research-administration guidance, standards updates and independent tool reviews.

Ask CASRAI · free to try

Ask about OMFIT Fusion Modeling Framework: Access, Modules, Use

2 free questions a day, no account.

An AI assistant for research: funder rules, compliance, methods and publishing. It links the sources behind its answers and labels answers that come from a web search.

Answers draw on CASRAI's guides and dictionary plus the federal and funder documents we index: Federal Register, Grants.gov, Regulations.gov and UKRI.

Works on this site and inside Claude, Cursor and the AI tools you already use.

Everything CASRAI publishes — this page, the dictionary, the guides and the news — stays open to read, with no account and no card.

Referenced across the research world

University of Cambridge logoColumbia University logoCrossref logoUniversity of Edinburgh logoHarvard University logoUniversity of Oxford logoPrinceton University logoStanford School of Medicine logoUniversity College London logoORCID logoUniversity of Cambridge logoColumbia University logoCrossref logoUniversity of Edinburgh logoHarvard University logoUniversity of Oxford logoPrinceton University logoStanford School of Medicine logoUniversity College London logoORCID logo
  • University of Cambridge logo
  • Columbia University logo
  • Crossref logo
  • University of Edinburgh logo
  • Harvard University logo
  • University of Oxford logo
  • Princeton University logo
  • Stanford School of Medicine logo
  • University College London logo
  • ORCID logo

View CASRAI adoption →