Molniya
Ergonomic data analysis for TypeScript.
A simple, typed DataFrame library. In-memory, zero dependencies, and blazing fast with SIMD line finding.
bun add molniya Everything you need
A complete toolset for data manipulation and analysis, built for the modern TypeScript ecosystem.
Bulletproof Type Safety
Molniya leverages TypeScript's advanced type system to infer schemas directly from your data. Get full autocomplete and compile-time checks for your column names and types.
// df.col("invalid") ➔ Error!
// df.col("price").sum() ➔ Correctly typed
Blazing Fast I/O
SIMD-accelerated line finding and direct byte parsing. 7.3M rows in 1.2s.
Familiar API
Pandas/Polars inspired syntax. filter, groupby, and select just work.
Lazy Frames
Work with datasets larger than RAM. Process data in streams without loading the whole file.
Zero Deps
Lightweight and portable. No heavy C++ binaries or complex native dependencies.
Advanced Aggregations
Robust GroupBy engine with support for sum, mean, std, and custom collectors.
"price": "mean",
"quantity": "sum"
});
Designed for Developers
A familiar API that feels like Pandas, powered by a high-performance engine designed for the TypeScript ecosystem.
import { readCsv } from 'molniya';
// Blazing fast SIMD-accelerated reading
const { df } = await readCsv('bitcoin_7m_rows.csv', {
delimiter: ',',
hasHeader: true
});
console.log(df.shape); // [7381118, 8]
console.log(df.head(5));Real-World Performance
Loading 387MB (7.3M rows) Bitcoin CSV dataset.
* Benchmarks performed on M1 MacBook Air. Arquero and Danfo run on Node.js using their standard CSV utilities.