v0.3.1 New release with Metis MICOM• Tark update• MGA MGA

A desktop workspace for PLS-SEM models, powered by seminr

Build, analyse, and report PLS-SEM models through a focused desktop workflow designed for researchers.

Download for Windows See how it works
Play Video

One workflow for model design, calculation, and interpretation

metis connects the visual model, your dataset, the seminr engine, and the results flow in one readable desktop environment.

metis, TAM_Study.metis React + TS Save Export ↩ ↪ ▶ Calculate Select LV Connect VARIABLES · PEOU_1 · PEOU_2 · PU_1 · PU_2 · ATT_1 · BI_1 · BI_2 PEOU_1 PEOU_2 PEOU_3 PU_1 PU_2 ATT_1 ATT_2 BI_1 BI_2 β=0.41** β=0.28* β=0.57*** PEOU Ease of Use PU Usefulness ATT Attitude BI Intention ▶ running PLS-SEM… React + Electron IPC / REST POST /calculate { model:{ constructs, paths, file } 200 OK { paths:{ PEOU_PU:0.431 PU_BI:0.614 ATT_BI:0.341 srmr:0.058 } localhost:8000 plumber.R model.R R 4.3.2 1 #* @post /calculate 2 function (model) { 3 # load CSV from model spec 4 dataset <- read.csv(model$file) 5 mm <- constructs (model$constructs) 6 sm <- relationships (model$paths) 7 8 pls <- estimate_pls ( 9 data = dataset, 10 measurement_model = mm, 11 structural_model = sm 12 ) 13 summarise_pls (pls) 14 } R Console seminr v3.2.1 > estimate_pls(data, mm, sm) ✓ PLS estimation complete β PEOU→PU 0.431 (p<0.01) β PU→BI 0.614 (p<0.001) β ATT→BI 0.341 (p<0.01) R²(BI) = 0.61 ✓ SRMR = 0.058 · NFI = 0.921 · AVE all ≥ 0.50 > _
about metis

An open desktop environment for serious PLS-SEM work

metis is a free desktop application for researchers and students who need a serious PLS-SEM workflow without expensive licences. It combines a visual modelling canvas with the R and seminr analysis engine so you can move from theory to interpretable results in one place.

visual model design

Build constructs and structural paths on an interactive canvas, then keep coefficients and model outputs connected to the diagram that produced them.

complete PLS-SEM workflow

Run PLS-SEM, faster bootstrap defaults, PLS predict, and higher-order construct workflows with practical control over subsamples, weighting schemes, and iterations.

no-code analysis workflow

Use seminr-powered statistics without writing R code. metis handles the backend pipeline while you stay focused on theory and interpretation.

private local computation

Your dataset and calculations stay on your machine. metis runs the analysis engine locally, with no cloud upload requirement.

canvas

Keep the model and the results in view

Move from the canvas to coefficients and back without losing context.

metis keeps path coefficients, loadings, and fit indices close to the model that produced them, so the structure and the interpretation stay aligned.

PATH COEFFICIENTS Path Coeff. 2.5% 97.5% p PEOU → PU 0.431 0.284 0.572 0.002 PU → BI 0.614 0.491 0.738 0.000 PEOU → BI 0.183 0.041 0.329 0.018 FIT INDICES SRMR 0.058 NFI 0.921 R² (PU) 0.61 R² (BI) 0.53
DATA, TAM_Study.csv peou1 peou2 peou3 pu1 pu2 5 4 5 4 3 3 3 4 5 5 4 5 4 3 4 INDICATOR ASSIGNMENT PEOU → peou1–4 PU → pu1–4 BI → bi1–3 248 rows · 12 columns · no missing values ready to model
data

Work with real datasets from the start

Build with your files, not placeholder examples.

Import your dataset and design with real indicator names, real loadings, and real significance values from the first run onward.

zero-code

No R required

Let the statistical engine run in the background while you stay with the research.

metis handles bootstrapping, PLS predict, NCA, cIPMA, IPMA, and the seminr setup work for you, so you can focus on specification, interpretation, and reporting.

BOOTSTRAP SETTINGS Subsamples 1000 change Inner Weighting path weighting centroid factor Max Iterations 300 run bootstrap
localhost:6000 no cloud. no account. no telemetry. R engine · local REST API · Windows + macOS
privacy

100% local processing

Your data never leaves your machine.

The R engine runs on localhost only. No account required, no telemetry, no cloud. Your research data stays yours.

metis against established SEM tools

Workflow
metis
SmartPLS
ADANCO
AMOS
Free local desktop workflow
-
-
-
Pricing & Subscription
Free forever
Paid
Paid
Paid
PLS-SEM focus
-
Visual model builder
Easy HOC drawing
-
-
PLSpredict workflow
-
-
IPMA and NCA
-
-
Tark-style report tables
-
-
-
See HOC docs Read the 0.3.1 update
tark

Journal-ready tables from saved results

Report Setup

Select Model & Workspace

Choose your target workspace, saved PLS-SEM model, and set custom report titles.

Tark Step 1 Report Setup Dark Tark Step 1 Report Setup Light
Path Diagram

Configure Path Values

Toggle structural path coefficients, outer loadings, and R² construct indicators.

Tark Step 2 Path Diagram Dark Tark Step 2 Path Diagram Light
Word Document

Specify Save Destination

Define output file name, destination directory, and review report summaries.

Tark Step 3 Word Document Dark Tark Step 3 Word Document Light
Journal Export

Export Journal-Ready .docx Tables

3 simple steps to generate publication-standard Word tables with formatted APA notes.

teaching_and_social_Tark_report.docx APA 7 Format
Table 1. Measurement Model & Construct Reliability
Construct Indicator Loadings CR AVE
PEOU 0.81 / 0.84 / 0.79 0.91 0.68
PU 0.83 / 0.85 / 0.80 0.90 0.67
BI 0.88 / 0.86 0.92 0.74
3 Steps Complete • Ready for thesis & journal publication
micom beta

Measurement Invariance for Composite Models

Verify measurement invariance across groups with the 3-step MICOM procedure before conducting multigroup analysis.

Step 1

Configural Invariance

Qualitative check confirming identical indicator assignments, composite weighting modes (Mode A/B), data treatment, and estimation algorithms across groups.

Automatic Pre-check
Step 2

Compositional Invariance

Permutation testing of composite weight correlation c (1,000 to 5,000 draws). Includes deterministic sign alignment algorithm to prevent false non-invariance.

Permutation Test (p ≥ 0.05)
Step 3

Equal Means & Variances

Evaluates composite score mean equality and log-variance equality across pooled sub-populations to establish full measurement invariance.

Full Invariance Verification
Read MICOM Technical Paper & Guide
multigroup analysis

Multigroup Analysis (MGA)

Compare structural relationships across distinct sub-populations, user demographics, or experimental conditions.

Henseler's MGA & PLS-MGA

Non-parametric permutation and bootstrap tests to determine if path coefficients differ significantly between group subsets (p < 0.05 or p > 0.95).

Parametric Welch & Pooled t-Tests

Evaluate path differences assuming equal or unequal variances across sample sizes, providing robust statistical verification.

Direct Tark Report Export

Automatically generate APA 7 formatted MGA path comparison tables, confidence intervals, and effect sizes ready for publication.

See MGA documentation
wall

notes on the wall

Notes from people using, testing, and shaping metis as an open-source research tool. Drag a card; press a corner; the cards push back.

see the full wall
FAQ

Quick answers

What is Metis?

Metis is a desktop application for Partial Least Squares Structural Equation Modelling (PLS-SEM). It brings dataset management, visual model building, statistical analysis, results exploration, and reporting into one workspace. Statistical calculations run through R, primarily using seminr, while Metis provides the visual workflow around the analysis.

Which installer should I choose: Bundle or Lite?

Both installers provide the same Metis application. Bundle is recommended for most users—it includes the complete R calculation environment and does not require installing R separately. Lite is intended for users who already have R installed.

Do I need to know R to use Metis?

No. Metis uses R in the background, but normal model building, analysis and reporting are performed through the graphical interface. Users who want to inspect or reproduce their analysis in R can also export or copy the corresponding R script.

Does my dataset leave my computer?

No. Model data, calculations, workspaces, saved results and Tark reports are processed and stored locally on your computer. Metis does not require your research dataset to be uploaded to a cloud analysis service.

Is Metis free?

Yes. Metis is distributed free of charge and released under the GNU General Public Licence v3 (GPL-3). There is no subscription required to run analyses.

What analyses are currently available in Metis?

The workflow includes PLS-SEM, Bootstrapping, PLSpredict, CVPAT, Importance-Performance Map Analysis (IPMA), Necessary Condition Analysis (NCA), combined cIPMA, 3-stage MICOM measurement invariance, Multigroup Analysis (MGA), higher-order constructs, moderation analysis, and automated Tark reporting.

What is Tark?

Tark is the reporting assistant built into Metis. It collects selected saved results and converts them into structured tables and optional model diagrams for use in a thesis, dissertation, paper or report, significantly reducing repetitive formatting work.

Can I use Metis for a thesis, dissertation or journal article?

Yes. Metis is designed for academic research workflows. For the current release, cite Metis in-text as (Metis, 2026) and seminr as (Ringle et al., 2024).

Explore all FAQs

Choose your download

A research-grade desktop workflow for building and interpreting PLS-SEM models.

Start with the version that fits your setup and keep model design, calculation, and interpretation in one connected environment.

Bundle 283 MB+

metis Bundle

  • Includes R runtime support, no separate R setup needed
  • Fully self-contained, works offline
  • seminr pre-installed and configured
  • Recommended for most researchers
View Bundle downloads
Lite 78 MB+

metis Lite

  • Smaller download, faster to install
  • Ideal for users already running R
  • seminr installed on first launch
  • Requires R to be installed separately
View Lite downloads
How to cite

Cite metis in your work

If metis supports your thesis, paper, teaching, or research project, please cite the software.

In-text citation (Metis, 2026)
Reference Metis. (2026). Metis: Free PLS-SEM desktop software for academic researchers (Version 0.3.1) [Computer software]. https://metis.emend.it.com
View full citation formats
user guide

metis User Guide

Explore user guides, workflow tutorials, and analysis documentation.

read the docs
August 2026

v0.3.1

Preferences and algorithm defaults hub, interactive missing-value finder, modernized single-file workspaces, higher-order construct modeling, and enhanced predictive summaries.

release notes
academic roots

Developed in an academic research environment

Built with early feedback from researchers at KNUST and now opening to the broader international academic community.