Trama
ITEN
Download

From data to figure. In a single project.

Trama brings worksheets, statistical analyses and scientific graphs together in one Windows application. Paste from Excel, pick a method, refine the figure and export to SVG or PDF. It is free, open source and needs no installation.

Beta 0.95.0 MIT licence · Windows 10/11, 64-bit · Vetro workspace in Italian and English · multicore calculation. Errors are possible: see the limits.

Dose–response · Trama 0.95 beta
Trama’s dark Vetro workspace with three fitted dose–response curves.
  • Linked sheetsSwitch between measurements, analyses and figures in the same project.
  • Dose–response curvesCompare three fitted curves with the observations on a log scale.
  • Vetro workspaceUse light or dark mode, or follow the system setting.

Trama 0.95 on Windows. Dose–response study with synthetic data.

  • €0MIT licence, professional use included
  • Qt 6Vetro workspace
  • 118catalogue equations in 14 families
  • 0network modules in the source

Where Trama works differently

We do not claim “more methods than anyone”: coverage is still partial, and we say so below. These are the points where Trama works differently from traditional commercial software, and where to check each one.

  • Free, no licences to manage

    No cost, subscription or seat count. MIT licence: use it, study it and modify it, including in professional work.

    Check: LICENSE
  • It just starts

    Extract the ZIP folder and run Trama.exe. No Python to install, no account to create, no activation key to enter.

    Check: getting started (it)
  • Your data stays on your computer

    The application source imports no network module: no telemetry, no sync, no online licence check.

    Check: search urllib, socket
  • A readable project file

    In the simple case a .alab project is JSON text: open it in any editor, compare it with a diff, archive it without depending on the program.

    Check: sample file
  • Method and limits written into the result

    An exact test says when it switched to an approximation and why. An interval that cannot be computed is explained, not filled in. The robust fit is descriptive: no invented standard errors or p-values. Analysis from summaries does not reconstruct artificial observations.

    Check: features (it)
  • Data, analyses and graphs stay linked

    Edit a value and the linked analyses recalculate. An excluded cell stays in the sheet and is left out of the calculation, not deleted.

    Check: first experiment (it)
  • Vector figures for papers and posters

    Export to SVG, PDF (also multipage), PNG or TIFF. In native layouts, up to 12 panels, compatible vector parts stay vector; raster parts stay images.

    Check: graphs and figures (it)
  • Public verification, limits included

    Verification reports record the test environment, test outcomes and portable ZIP checksum. The link below is the 0.94 report.

    0.94 verification

Trama and traditional commercial software

The comparison is about how you work, not about a list of methods. The right-hand column describes what often happens: it does not refer to any particular product.

AspectTramaTraditional commercial software, usually
CostFree. MIT licence, professional use included.A licence or subscription, per user or per seat.
Getting startedPortable folder: extract and run. No account, no activation key.An installer; in some cases activation or an account sign-in.
Code and calculationsPublic source, built on open libraries (NumPy, SciPy, statsmodels). You can read how a result is computed.Closed code: results are checked by comparing them with other tools, not by reading the implementation.
Project fileIn the simple case, JSON text that any editor can open.Often a proprietary format that only the program that created it can read.
DataLocal files. The source contains no network code.Varies by product: sync and data collection need checking case by case.
ContinuityWith the MIT licence the code stays usable and modifiable even if the project stops.Depends on the vendor: an expired licence or end of support can limit access to your own projects.

Terms differ between vendors and change over time: check those of the program you use today. Trama does not claim equivalence with other programs or general superiority; the timing measurements in the repository describe synthetic workloads only and are not comparisons.

Real screenshots

Real captures of the program, with synthetic data.

What’s new in 0.95 · The new Vetro interface uses Qt 6 with PySide6. Keep projects open in tabs, browse graphs in the Gallery and draw directly on figures.

Multicore calculation. Multicore calculation can spread work across processes for selected linked analyses, K-means, Monte Carlo studies, ROC bootstrap and paired permutations. Choose Automatic or set the number of processes in Settings → Multicore calculation. Speed gains depend on the calculation and the size of the dataset. In a Windows 11 VM with 10 logical CPUs and workers already running, 10,000 Monte Carlo replicates took 84 ms instead of 281 ms (3.32×), and ROC bootstrap took 0.53 s instead of 1.30 s (2.48×). These are medians of seven runs, with identical results from 1 to 8 processes. Benchmark data.

Gallery
Gallery with five ready chart and layout previews.
Open a figure from its preview.Find charts and layouts by name and jump to the sheet you need.
Treatment distributions
A violin plot for three treatment groups in the light workspace.
See how the groups differ.Show the shape and range of each group with violin plots.
Graph Inspector
Dose–response chart with the Graph Inspector on the right.
Refine the figure in the Inspector.Preview axes, labels and colours while you edit.
Response profile
Dose–response curves and treatment distributions in a two-panel layout.
Compose the figure for export.Arrange panels and export them as vector SVG or PDF.

How these screenshots were captured.

What you can do

Six families of tools in 0.95.0 beta, from the same worksheet. Open the full map of formats and features.

Compare experiments

t-tests (also lognormal and TOST equivalence), exact rank tests, ordinary, Welch and Brown–Forsythe ANOVA, repeated measures, REML mixed models with Kenward–Roger, multiple comparisons and compact letters on the graph.

Fit and compare curves

Asymmetric parameter profiles; 118 catalogue equations in 14 families, custom and conditional equations with fixed thresholds, global fits with shared parameters, weighting, AICc comparison, direct and inverse calibration, splines, LOWESS and robust fits.

Explore complex data

PCA with biplot, Lloyd and Hartigan–Wong K-means with a 17-index consensus and reports for each k, hierarchical clustering with dendrograms, merge maps and WCSS elbow plots, correlation matrices, heat maps and confidence ellipses.

Study outcomes and associations

Kaplan–Meier, Cox and competing risks, ROC curves with DeLong comparison, Bland–Altman, multiple, logistic and Poisson regression, contingency tables and diagnostic measures.

Prepare observations

In-sheet formulas, transformations, baselines also with replicates, normalisation, area under the curve and peaks, Monte Carlo studies with a reproducible seed.

Compose the figure

XY and column graphs, box and violin plots, forest and bubble plots, asymmetric error bars, native annotations and shapes, layouts up to 12 panels.

Full list, options and limits in the feature overview (Italian).

What Trama is not (yet)

Better to know before you choose.

  • It is a beta. 0.x versions can contain errors and incomplete features; there is no stable release yet.
  • Partial coverage. We do not claim equivalence with other programs: an established package may offer options, formats, official support and years of maturity that Trama does not have.
  • Windows 10/11 64-bit only for the portable package.
  • Importing projects from other programs is limited to verified content; there is no universal export. See compatibility (Italian).
  • The interface is available in Italian and English, but some result tables, legends and messages remain in English.
  • No clinical validation or regulatory certification. Tests prove the cases that were run, not every possible use.

Frequently asked questions

Short answers to the questions we hear most often.

Is Trama really free?
Yes. Trama is released under the MIT license: no cost, subscription or per-seat limit, including for professional use.
Which systems does it run on and how do I install it?
The portable package is for 64-bit Windows 10 and 11. Extract the ZIP folder and run Trama.exe: no Python, account or activation key is needed. You need a screen of at least 1024 × 768 pixels.
Which statistical analyses and graphs does it offer?
t tests (including lognormal and equivalence TOST), exact rank tests, ANOVA, repeated measures and mixed models, linear, logistic and nonlinear regression with 118 catalogued equations, global fits, Kaplan–Meier, Cox and competing risks, ROC curves, PCA, K-means and hierarchical clustering. Graphs include XY, columns, box and violin plots, heat maps, forest and bubble plots, exported as SVG, PDF, PNG and TIFF.
Can I import my data from Excel or CSV?
Yes: you can paste from Excel and import CSV, TSV, TXT and XLSX files (sheets are imported as data). The native .alab project is JSON text that opens in any editor.
Does Trama collect data or need an internet connection?
No. The application source has no network modules: no telemetry, no sync and no online license check. Your data stays on your computer.
Can I use it for clinical or regulatory results?
Trama is a beta with partial coverage and no clinical validation or regulatory certification. Always verify the results; the limits are written on the page and in the release notes.
Where is the source code?
On GitHub, in the realfulvio/Trama repository, together with the quick guide, the feature list and the release history.

Try it in three steps

About five minutes, nothing to install. Needs Windows 10 or 11, 64-bit, and a screen of at least 1024 × 768 pixels.

  1. Download

    Get Trama-<version>-Windows-x64.zip from the releases page, not the source archive.

  2. Extract and run

    Extract the whole folder, keep the _internal folder next to the executable and run Trama.exe.

  3. Open the sample

    Download Laboratorio.alab and open it: synthetic data, a global fit and a ready-made layout.

Ready to try it on your own data?

Free, portable, in Italian and English.