A calmer way to build and read
PLS-SEM models.

metis brings model design, calculation, and interpretation into one research-grade desktop workflow, so you can follow the logic of the analysis from structure to result.

See how it works
metis — TAM_Study.metis
TAM_Study.csv
100 cases · 12 vars
search…
variables
PEOU_1
PEOU_2
PEOU_3
PEOU_4
PU_1
PU_2
PU_3
PU_4
ATT_1
ATT_2
BI_1
BI_2
PEOU_1 PEOU_2 PEOU_3 PEOU_4 PU_1 PU_2 PU_3 PU_4 ATT_1 ATT_2 ATT_3 ATT_4 BI_1 BI_2 BI_3 BI_4 β=0.41** β=0.28* β=0.57*** β=0.34** PEOU Ease of Use PU Usefulness ATT Attitude BI Beh. Intention ▶ running PLS-SEM · seminr v3.2.1…

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.

R
R
seminr
CSV
CSV/Excel
installer
Electron
TS
TypeScript
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, bootstrap, and PLS predict 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, and the seminr setup work for you, so you can focus on specification, interpretation, and reporting.

seminr runs in the background
The R engine computes locally while you work through the interface.
Keep control of the method
Adjust bootstrap subsamples, inner weighting, and algorithm settings without leaving the workflow.
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 only
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.

testimonials

Trusted by researchers

Early feedback from lecturers, researchers, and postgraduate users is helping shape the public beta.

"

metis changed how I teach PLS-SEM to postgraduate students. The visual canvas makes abstract structural modelling concepts immediately tangible.

DA
Dr. Adomako
Researcher, KNUST
"

Finally a tool that connects the model I draw on the whiteboard with the actual seminr computation. Nothing was this seamless before metis.

FK
Dr. Francis Kofi Nunoo
Lecturer, KNUST
"

I used metis for a multi-group analysis and the setup that would have taken days in other tools was complete in a single afternoon. Remarkable efficiency.

GO
George Owusu
PhD Researcher, KNUST
"

The visual canvas removed every barrier I had with SEM. I ran my first bootstrapped model in under an hour — no code, no confusion. metis is a game changer.

PA
Prosper Asante
Graduate Researcher
"

My dissertation committee was impressed by the path diagram output directly from metis. It made the whole methodology section look highly professional.

RN
Raphael Nii Okai
MPhil Student
"

The fit indices update as I refine the model. That live feedback loop during model specification is invaluable. A genuine research-grade tool.

BG
Dr. Boakye
Academic, KNUST
Our Team 6

The People Behind
metis

Our team brings together research leadership, engineering, design, and UX practice to build a clearer PLS-SEM experience. We share one goal: to make serious structural modelling more accessible without diluting the method.

View all team

Choose your beta build

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 ~264 MB

metis Bundle

  • Includes R Portable — no R installation needed
  • Fully self-contained, works offline
  • seminr pre-installed and configured
  • Recommended for most researchers
Lite ~72 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
roadmap

metis Roadmap

See what is coming next and where the workflow is heading.

view the roadmap
April 2026

v0.1.7

Visual canvas, PLS-SEM, bootstrap resampling, and PLS predict

build log
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.