Benchmark Comparison of Packages with Runtime Validation and TypeScript Support
A benchmark suite comparing the runtime performance of dozens of TypeScript data-validation and parsing libraries, such as zod, valibot, ArkType, and io-ts, run in isolated Node.js processes.
This project benchmarks the runtime performance of a large set of JavaScript and TypeScript data-validation packages, including zod, valibot, ArkType, io-ts, superstruct, typebox, yup, and dozens of others. Each package is measured on validating data for type correctness and on providing either a validator function that returns a cast value or throws, or a type-guard function that narrows a value's type.
Benchmarks run in isolated Node.js processes per package to keep results independent, and can also run under Bun. Commands are provided to run the full suite or a chosen subset of packages, and a small web viewer displays the results.
The project also includes a build and test pipeline that regenerates the benchmarked code for each package it measures.