breadroll is a lightweight TypeScript library for type-safe data processing. It is optimized for Bun and designed to make common data workflows feel safer and more pleasant inside TypeScript or JavaScript projects.

The library centers on a dataframe-style workflow. You can open local or remote data sources, read formats like CSV, TSV, and JSON, select columns, inspect rows, and perform transformations with a small API. The goal is to make everyday data manipulation feel approachable without pulling in a large data science stack.

The project is especially focused on developer experience. Instead of treating data processing as an untyped pile of arrays and objects, breadroll leans into TypeScript so common operations can be easier to reason about while still feeling flexible. It tries to keep the API simple enough for quick scripts but structured enough for repeatable workflows.

Because breadroll is built with Bun in mind, it also explores what modern JavaScript runtimes make possible for data tooling. Bun’s file APIs and performance characteristics make it a good fit for reading local files, loading remote data, and running fast transformations in a lightweight package.

The project is open source and doubles as a space for exploring typed data utilities, package design, runtime-specific optimizations, and practical data transformation tools for JavaScript developers.

I built breadroll as both a useful utility and an experiment in how data tooling can feel in the TypeScript ecosystem: small, typed, fast to install, and pleasant to reach for when a project needs to parse, inspect, or reshape structured data.