A modular toolkit for detecting AI-assisted responses in crowdsourced behavioral research.

slopit provides tools to capture and analyze behavioral signals during text entry tasks, helping researchers identify potentially AI-assisted responses in crowdsourced studies.

TypeScript packages for client-side behavioral capture in web experiments:

• Keystroke dynamics (timing, patterns)
• Paste event detection
• Focus/blur tracking
• Platform adapters for jsPsych, lab.js, PsychoJS, Gorilla, JATOS, OSWeb, PCIbex

Python package for server-side analysis:

• Behavioral feature extraction
• Stylometric analysis
• Machine learning classifiers

• TypeScript Documentation: Behavioral capture packages
• Python Documentation: Analysis toolkit

npm install @slopit/behavioral @slopit/adapter-jspsych

import { SlopitExtension } from "@slopit/adapter-jspsych";

const jsPsych = initJsPsych({
  extensions: [{ type: SlopitExtension }],
});

pip install slopit

from slopit import Session

session = Session.from_json(data)
features = session.extract_features()

• Node.js >= 22
• pnpm >= 9
• Python >= 3.13
• uv (Python package manager)

# Install Node dependencies
pnpm install

# Install Python dependencies
cd python && uv sync --extra dev

# Build all packages
pnpm build

# Run tests
pnpm test

# Type check
pnpm typecheck

# Lint
pnpm lint

See CONTRIBUTING.md for development guidelines.

MIT

This project was architected and implemented with the help of Claude Code.