react-native-uuid is a zero-dependency TypeScript implementation of RFC4122 standard A Universally Unique IDentifier (UUID) URN Namespace. Please note, this library uses pseudo random generator based on top of Math.random. New version with hardware support is WIP.
Heavily inspired by:
Huge thanks to Randy Coulman for the early version of a code.
- Node.js: 18.0.0 or higher (16.x also supported)
- npm: 9.0.0 or higher
npm install react-native-uuidimport uuid from 'react-native-uuid';
uuid.v4(); // ⇨ '11edc52b-2918-4d71-9058-f7285e29d894'# Install dependencies
npm install
# Run tests
npm test
npm test -- --watch # Watch mode
npm test -- --coverage # With coverage report
# Linting and formatting
npm run lint # Lint TypeScript and JavaScript
npm run prettier:write # Format all files
npm run prettier:check # Check formatting without changes
# Building
npm run build # Build TypeScript to JavaScript
npm run prepublishOnly # Build before publishing (auto on npm publish)
# Benchmarking
npm run bench:rn # Generate benchmark matrix for RN 0.71 - 0.77
npm run bench:security # Run security benchmark with pass/fail thresholds
# Documentation
npm run docs # Generate TypeDoc documentationWe keep a baseline benchmark matrix for major React Native versions:
- React Native: 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77
- Cases:
v1,v3,v4,v5,validate,parse,unparse - Stored benchmark results:
benchmarks/results/rn-version-benchmark.md - Stored benchmark data:
benchmarks/results/rn-version-benchmark.json - Stored test report:
benchmarks/results/test-results.json
Generate fresh results locally:
npm run bench:rn
npm test -- --runInBand --json --outputFile benchmarks/results/test-results.json
npm run bench:security
npm run bench:figures
The security benchmark is designed as a pass/fail gate for randomness quality and misuse safety checks.
- Command:
npm run bench:security - Stored report:
benchmarks/results/security-benchmark.md - Stored data:
benchmarks/results/security-benchmark.json - CI workflow:
.github/workflows/security-benchmark.yml
Current threshold gates:
- v4 sample collisions:
0 - Cross-batch overlap (restart proxy):
0 - Monobit test z-score:
<= 4 - Runs test z-score:
<= 4 - RNG chi-square score (256 bins):
<= 420 - Malformed input crashes:
0
The CI job fails if any threshold is exceeded.
Methods documentation is available here
Previous version has been based on randombytes that is not compatible with react-native out of the box.
Please submit an issue if you found a bug.
See the contributing guide to learn how to contribute to the repository and the development workflow.
Thank you to our sponsors:
Copyright (c) 2016-2026 Eugene Hauptmann