Free CB-SEM software, just draw your model

Yes. Every feature is free, with no account, no trial period and no limits on sample size or model size. It runs entirely in your web browser.

No. Your data file is read and analysed inside your browser tab. Nothing is sent to our server, so confidential survey data stays on your computer.

For the standard CB-SEM workflow, yes: you draw the model as you would in AMOS, estimate it by maximum likelihood, and get model fit, standardized and unstandardized estimates, R², indirect effects and modification indices. Commercial packages offer more, such as multigroup invariance testing, Bayesian SEM and full information maximum likelihood for missing data. Results follow the same maximum likelihood method as lavaan and AMOS; the modification indices reproduce lavaan’s published values.

CB-SEM estimates common factors and tests how well the whole model reproduces the observed covariance matrix, so it reports global fit indices such as CFI, TLI and RMSEA. PLS-SEM builds composites and maximises explained variance, which suits prediction, formative constructs and small samples. Use CB-SEM to confirm a theory; use the free PLS-SEM Studio when prediction is the goal.

Chi-square with degrees of freedom and p value, normed chi-square (χ²/df), CFI, TLI, RMSEA with its 90% confidence interval and PCLOSE, SRMR, GFI, AIC and BIC, each flagged against the common cut-offs of Hu and Bentler (1999).

Draw the paths, for example X → M → Y and X → Y. Every indirect route is found automatically and tested with a Monte Carlo confidence interval (Preacher and Selig, 2012) as well as the delta-method standard error, and the result is classified as full, partial or no mediation (Zhao, Lynch and Chen, 2010).

In the Moderation box of the model builder, choose the moderator and the path it changes. CB-SEM Studio builds a latent interaction from double-mean-centred matched-pair product indicators (Marsh, Wen and Hau, 2004; Lin et al., 2010), adds the moderator’s main effect, and reports the interaction effect, simple slopes at ±1 SD with Monte Carlo confidence intervals, a simple slope plot and, where a moderated path is part of an indirect route, the index of moderated mediation.

A modification index estimates how much the chi-square would drop if a fixed parameter, such as a residual covariance or a cross-loading, were freed. CB-SEM Studio lists them with the expected parameter change, and residual covariances between items of the same construct can be added with one click. Only free parameters that make theoretical sense, and report every change.

Yes. For each construct you get standardized loadings, Cronbach’s alpha, composite reliability (CR), average variance extracted (AVE), maximum shared variance (MSV), the HTMT ratio and the Fornell–Larcker criterion, each flagged against common thresholds (Hair et al., 2019; Henseler et al., 2015).

Import CSV, TSV, TXT, XLSX, XLS and ODS files. Export every table to Excel, Word, HTML, CSV and LaTeX, the full results to JSON, the model as lavaan syntax, and the path diagram as SVG or 300 dpi PNG.

Please cite the software whenever you use it in research: Xavier, B. (2026). CB-SEM Studio: Free covariance-based structural equation modeling software [Web application]. FreeStatisticalTools.com. https://freestatisticaltools.com/cb-sem/ A BibTeX entry is in the How to cite section of this page.

CB-SEM Studio was created by Bibin Xavier as part of FreeStatisticalTools.com, together with the free PLS-SEM Studio and more than 90 other statistical tools.