Research tools that run in your browser.

Plot, test, fit, convert, inspect structures and tidy citations: 18 research tools, free and open source. Files you open are read in your browser and never uploaded.

Data & Statistics

8 tools

Clean, test, fit and plot experimental data.

Plot Digitizer

Extract data points from a graph image by clicking or tracing curves, then export them as CSV, a replotted figure or Python.

Data Cleaner

Clean messy CSV files step by step in an editable recipe, then download the tidy file and the pandas script that repeats it.

Statistics Calculator

Run t-tests, ANOVA with post-hoc comparisons, correlation and non-parametric tests, with effect sizes, plots, advice and a Python script that reruns them.

Error Bar Generator

Compute mean, SD, SEM, median, IQR and confidence intervals per group, plot error bars with significance brackets, and get the Python.

Outlier Detector

Flag anomalies with Z-scores, Tukey’s IQR fences or the robust modified Z-score, see them plotted, and get the Python.

Curve Fitter

Type any equation and fit it to your data: parameters with uncertainties, a styled plot, and the Python code.

Plot Builder

Plot CSV columns as lines, bars, histograms, box plots, heatmaps and more, on one or several panels; style it and export PDF, PNG, SVG or the matplotlib script.

XVG Visualizer

Plot GROMACS .xvg and CSV logs: pick columns, add running means and panels, and export the figure or the Python that draws it.

Computational Chemistry

4 tools

Inspect structures, prepare coordinates and write the batch scripts for a run.

Structure Inspector

Render PDB, CIF, GRO, XYZ and more in 3D; select atoms, measure geometry and export a figure.

Coordinate Manipulator

Translate, rotate and re-box PDB, GRO and XYZ coordinates, and convert between the formats.

Scientific Converter

Convert energy, length, pressure, dipole, polarizability, spectroscopic and temperature units against CODATA values.

MD Workflow Generator

Set up GROMACS runs with every .mdp file explained, index groups and one zip; batch scripts for SLURM, PBS, LSF or Grid Engine; LAMMPS and PLUMED too.

Writing & Citations

6 tools

Equations, tables and a bibliography that builds.

Equation Formatter

Build LaTeX equations and matrices visually with a live preview, then copy the code.

Visual LaTeX Tables

Paste a table from Excel or CSV, preview it, and copy clean LaTeX (booktabs) or Markdown.

DOI to BibTeX

Paste DOIs and get ready-to-use BibTeX entries, looked up through doi.org.

BibTeX Deduplicator

Find entries in a .bib library that share a DOI or title, and keep one of each.

BibTeX Sanitizer

Tidy messy BibTeX exports: protect capitals in titles and clean up fields.

Journal Abbreviator

Convert full journal names to their ISO 4 abbreviations.

Focus & Flow

3 tools

Not research tools: a timer, a decision matrix and a physics toy for the breaks in between.

Ambient Pomodoro

A focus timer with adjustable ambient sounds like rain and cafe noise for deep work.

Decision Matrix

Weigh options against criteria in a scored matrix to rank your best choice.

Kinetic Sandbox

An interactive particle field that scatters from your pointer and gathers while you hold it, a calm visual break.

Use the library behind the tools

Most tools compute with stemkit-core: 42 domain modules covered by 4,434 tests validated against SciPy, NumPy and statsmodels. The same functions run in Node, so an analysis explored here can be scripted, version-pinned and re-run.

Install it from npm. It needs Node 18 or later and brings no other packages with it.
Terminal
```
npm install stemkit-core
```

Save an example in the same folder under the name shown, with any data file it reads beside it, and run it: node pick-test.mjs.

Summarise every series in a GROMACS .xvg file, named by its legend.

xvg-summary.mjs

import { readFileSync } from 'node:fs';
import { parseXvg, extractColumn, columnStats } from 'stemkit-core';

const { matrix, headers } = parseXvg(readFileSync('rmsd.xvg', 'utf8'));

headers.slice(1).forEach((name, i) => {
  const s = columnStats(extractColumn(matrix, i + 1));
  console.log(`${name}: ${s.mean.toFixed(3)} ± ${s.std.toFixed(3)}`);
});

Let the assumption checks pick the test, then run it.

pick-test.mjs

import { recommendTest, independentTTest, mannWhitneyU } from 'stemkit-core';

const control = [4.1, 4.3, 3.9, 4.4, 4.0, 4.2, 4.5, 3.8, 4.1, 4.6];
const treated = [5.0, 5.2, 4.8, 9.9, 5.1, 4.9, 5.3, 5.0, 12.4, 5.2];

const advice = recommendTest({
  design: 'independent', groups: [control, treated]
});
console.log(advice.reason);

const r = advice.test === 'mann-whitney'
  ? mannWhitneyU(control, treated)
  : independentTTest(control, treated);
console.log(`${advice.test}: p = ${r.p.toPrecision(3)}`);

// Group 2 departs from normality (D'Agostino p = .005), so a
// rank-based test is safer.
// mann-whitney: p = 0.000154

Mass and radius of gyration of a .gro structure, with the unit made explicit.

gyration.mjs

import { readFileSync } from 'node:fs';
import {
  parseStructure, centreOfMass, radiusOfGyration, convert
} from 'stemkit-core';

const gro = readFileSync('protein.gro', 'utf8');
const { atoms, unit } = parseStructure(gro, 'gro');
const { mass } = centreOfMass(atoms);
const rg = convert(radiusOfGyration(atoms), 'length', unit, 'angstrom');

console.log(`${atoms.length} atoms, ${mass.toFixed(1)} Da`);
console.log(`Rg = ${rg.toFixed(2)} Å`);

Convert between units on CODATA values.

units.mjs

import { convert, convertTemperature } from 'stemkit-core';

// 1 hartree in kcal/mol, 1 eV in kJ/mol, 298.15 K in °C
console.log(convert(1, 'energy', 'hartree', 'kcal'));  // 627.5094740631
console.log(convert(1, 'energy', 'ev', 'kj'));         // 96.48533212331287
console.log(convertTemperature(298.15, 'k', 'c'));     // 25

A SLURM array job for GROMACS, with a warning about memory in flight.

array-job.mjs

import { generateScript } from 'stemkit-core';

const { script, warnings } = generateScript({
  engine: 'gromacs', jobName: 'replicas', modules: ['gromacs/2023.3'],
  nodes: 1, gpus: 1, cpusPerTask: 16, memory: '32G', walltime: '24:00:00',
  array: true, arrayRange: '1-100%10', tpr: 'md.tpr', deffnm: 'md'
});

console.log(script);
warnings.forEach(w => console.warn(`[${w.level}] ${w.message}`));

// [warn] Array job: --mem=32G is per task. With 10 task(s) running
// concurrently that is 320G in flight. ...

Remove duplicate entries from a .bib library, keeping the most complete of each.

dedupe-bib.mjs

import { readFileSync, writeFileSync } from 'node:fs';
import {
  parseBibtex, deduplicateAuto, serialiseLibrary
} from 'stemkit-core';

const { entries } = parseBibtex(readFileSync('refs.bib', 'utf8'));
const { entries: kept, removed } = deduplicateAuto(entries);

writeFileSync('refs.clean.bib', serialiseLibrary(kept));
console.log(`${removed} duplicate(s) removed, ${kept.length} kept`);

Modules and worked examples are in the library reference.

What leaves your computer

Nothing you open. Files are read in the browser and never uploaded, there is no account, and a page keeps working offline once it has loaded.

Two features go online, and only when you use them: DOI to BibTeX looks up the DOIs you give it, and the structure inspector can fetch an entry from the PDB by its ID.

Cite STEMKit

This DOI always resolves to the current release.

10.5281/zenodo.21543112