Proactive step hints
One glance per tab
Open Results, Editor, MGA or Advanced and see 1 to 3 suggestions right at the top. Each hint reads your live model and cites your own numbers.
PLS-SEM · Open-source engine Structural equation modeling.
Structural equation modeling.
Finally modern.
The open platform for structural equation models: visual path editor, the full PLS-SEM pipeline and publication-ready reports. Right in the browser. Or self-hosted, when your data is sensitive.
Or start with the User Guide (76 pages, PDF)
No setup, no credit card · Open-source engine · Validated against established reference values
Live preview
A real model, fully computed.
This is what openpls-engine returns for a small Quality → Satisfaction → Loyalty chain. Paths, R², HTMT and bootstrap CIs in one view.
ECSI customer-satisfaction sample (n = 250), 5 000 bootstrap resamples. Reproducible in the demo project.
Workflow
Four steps from data to model.
- 01
Upload data
CSV, XLSX, SPSS or Stata: we read whatever you have.
- 02
Draw the model
Constructs and paths via drag & drop. Live preview of the model structure.
- 03
Compute
One click. plspm computes: loadings, paths, bootstrap confidences.
- 04
Export
Report as PDF, tables as XLSX, model as JSON or LaTeX snippet, all in open formats.
AI research companion New · Beta
Ambient AI that knows your project.
Contextual hints appear inline as you work. A chat drawer answers open questions about your model, all grounded in the numbers you already computed.
AI chat drawer
Ask, get grounded answers
Open the robot in the bottom right to ask anything about your model, data, or results. Four read-only tools give the model access to your project structure, never to raw rows.
Two designers, one connected flow.
Turn a research question into a starter model in seconds. Get a validated questionnaire back from the model's LVs. Collect data, then interpret it with the same assistant.
1
AI model designer
LiveType your research question. The assistant proposes 3–6 latent variables with modes (reflective / formative), hypothesised paths with citations, and 2–3 blueprint references from the PLS-SEM literature. One click writes it into a new model doc with LVs and paths pre-drawn.
Research question
"What drives customer loyalty in online grocery shopping?"
↓
Trust · Reflective Perceived Quality · Reflective Satisfaction · Reflective Loyalty · Reflective
2
Questionnaire builder
LiveGiven the model, the assistant generates 4–6 validated survey items per LV with source citations (Davis TAM, Zeithaml SERVQUAL, Cronin & Taylor …). Or start from scratch and write items yourself. Edit inline, add demographics, export as CSV / Markdown / JSON for Qualtrics, LimeSurvey or SoSci Survey.
Trust · Davis (1989)
trust1I feel confident this shop delivers on its promises.trust2This shop appears reliable in its transactions.trust3The shop takes reasonable care with my personal data. Complete toolkit New
Everything the literature asks for.
Fifteen modern PLS-SEM methods (MGA, IPMA, PLSpredict, HOC, FIMIX, PLSc, HTMT2, moderation, CVPAT, and more) in the same editor. No second tool. No plugin marketplace.
Quantify uncertainty in your effects and turn the model into out-of-sample predictions.
-
IPMA
Importance-Performance Map: surfaces constructs that are highly important and underperforming at the same time. Management priority at a glance.
-
PLSpredict
k-fold out-of-sample assessment with the full Shmueli panel: RMSE, MAE and MAPE versus a linear-model benchmark, plus an in-sample fit table for context.
-
Specific indirect effects
Every mediation chain in your model with point estimates and bootstrap confidence intervals, t-values and p-values. Stop reporting only the total indirect effect.
-
CVPAT (predictive ability)
Cross-validated predictive ability test after Liengaard et al. 2021. Head-to-head comparison against indicator-average or linear-model benchmarks, paired t-test, k-fold cross-validation.
-
FIMIX-PLS
Finite-mixture segmentation surfaces unobserved subgroups in your sample. Make heterogeneous effects visible.
-
MICOM
Three-step Henseler–Ringle–Sarstedt test for measurement invariance: configural, compositional and scalar. The required prerequisite before any multi-group comparison across countries, segments or time points.
Stress-test the measurement model and the structural equations against the most common biases.
-
PLSc (Consistent PLS)
Dijkstra-Henseler bias correction for reflective measurement. Per-LV ρ_A, dis-attenuated correlations, corrected paths and R².
-
Discriminant validity
HTMT plus the geometric-mean refinement HTMT2 (Roemer et al. 2021) and the Fornell-Larcker check side by side. Three independent perspectives on construct distinctness in one panel.
-
Gaussian copula
Park-Gupta / Hult et al. endogeneity test for structural predictors. Detects bias without instrumental variables, with built-in admissibility check.
-
Nomological validity
Pre-register expected signs for construct correlations and test each hypothesis with one-sided t-tests. Turns the theory chapter into a testable panel with a clean verdict per relationship.
-
Common method bias (Lindell-Whitney)
Marker-variable partial correlation adjustment to detect common method variance. Pick a theoretically unrelated marker, compare adjusted vs. raw correlations, decide whether CMB is a concern.
-
Moderation (Two-Stage)
Test interaction terms in your structural model. Two-stage approach after Henseler & Chin, in one click.
Build richer model topologies and pick the convergence algorithm that fits your data.
-
Higher-Order Constructs (HOC)
Disjoint two-stage workflow for hierarchical models. All four canonical types (R-R, R-F, F-R, F-F) plus nested HOCs, without rewiring the editor structure.
-
Newton scheme
Quasi-Newton inner weighting for more stable and faster convergence on complex models.
-
PCA scheme
Lohmöller’s PCA inner-weighting scheme as an alternative to centroid and path. Algorithmic choice per model.
New · Public surveys
From questionnaire to dataset. In one link.
Publish your questionnaire to a shareable URL, watch anonymous responses stream in, import as a dataset with one click. Everything researchers actually need built in.
Explore public surveys Version snapshots Public intro vs. internal notes Anti double-submit guard Owner-only response reads One-click dataset import Multi-locale respondent page
Methods, visualised
What the outputs actually look like.
Four examples of what OpenPLS produces for your analyses. Straight from the in-app explainer panels.
01
Bootstrap distribution
5,000 resamples, point estimate and 95 % CI. This is what the uncertainty around a path coefficient looks like.
02
Importance-performance map
Which constructs matter and underperform? A four-quadrant view for management prioritisation.
03
FIMIX segmentation
Find unobserved subgroups. Surface heterogeneous effects before they bias your results.
04
Simple slopes
How does a path coefficient change at low, mean and high moderator? Three lines say more than a single p-value.
Case Studies
Reproducible in 60 seconds.
Four published PLS-SEM models, available as one-click clones in your OpenPLS workspace. Each ships with synthetic original data, the complete path definition, and reproducible key metrics.
ECSI
European Customer Satisfaction Index
Customer Experience
Six reflective constructs (Image → Expectations → Quality → Value → Satisfaction → Loyalty), the canonical model from service marketing research.
- N
- 250
- LVs
- 6
- R²(SAT)
- 0.71
Fornell et al. (1996)
Open case study UTAUT2 + Trust
E-commerce acceptance with trust
Marketing / E-commerce
Performance expectancy, effort expectancy, social influence, and trust explain purchase intention and actual use. Includes demographics for multi-group analysis.
- N
- 400
- LVs
- 6
- MGA
- Age · Device
Venkatesh et al. (2012)
Open case study JD-R
Employee engagement (Job Demands-Resources)
HR / Organizational Behavior
Job demands and job resources affect satisfaction and performance via engagement. A classic mediator model with five constructs.
- N
- 350
- LVs
- 5
- R²(JP)
- 0.40
Bakker & Demerouti (2017)
Open case study TAM + Privacy
Adoption of digital health apps
Healthcare / mHealth
PEOU, usefulness, health consciousness, and privacy risk drive attitude and behavioral intention. A negative privacy path makes for a great f² effect-size example.
- N
- 380
- LVs
- 6
- Effect
- PR → BI negative
Davis (1989); Sun et al. (2013)
Open case studyCloning requires a free OpenPLS account.
Pricing
Open today. Open forever.
The web app is free to use. The engine is open-source under GPL-3.0 and always will be.
Cloud
Free
Everything you need for your research. Hosted by us.
- Unlimited models and datasets
- All reports & exports
- Visual path editor
- Version comparison
Self-host
$0 GPL-3.0
Engine as a Docker container and Python library. On your server, inside your network.
- Full engine functionality offline
- CLI and Python API
- Data sovereignty for clinic / industry
- Community support on GitHub
Consulting
On request Custom
Custom adaptations, training and methodological advice for research groups.
- Methodological coaching
- Custom models / plugins
- On-site workshops
- Priority support