Tiny-parquet – JavaScript lib to read/write Parquet in 326KB of WASM
Only Parquet lib fitting Cloudflare Workers free tier—flat schemas, 20MB/s throughput.

TagLib compiled to WASM fills a real gap—JS had no battle-tested audio tagging library.
JavaScript/TypeScript developers building music or media applications
ffmpeg.wasm · jsmediatags
This project exists because the TS|JS ecosystem had no battle-tested audio tagging library that supports reading and writing music metadata to all popular audio formats. TagLib-Wasm stands on the shoulders of giants, including TagLib itself, Emscripten, and Wasm (WebAssembly). TagLib itself is legendary, and a core dependency of many music/media apps.
import { readTags, updateTags } from "taglib-wasm/simple"; const tags = await readTags("song.flac"); console.log(tags.title, tags.artist, tags.album); await updateTags("song.flac", { title: "New Title", artist: "New Artist" });
On PC/server runtimes (Deno, Node.js, Bun), TagLib-Wasm auto-selects a WASI backend that does fast and efficient seek-based I/O. In browsers, it uses Emscripten with full-buffer I/O. No configuration is needed.This is the first 1.0 beta. I'd appreciate any feedback on the API design/DX, missing formats/features, and use cases I haven't thought of.
Only Parquet lib fitting Cloudflare Workers free tier—flat schemas, 20MB/s throughput.
Built-in reactive runtime and dual JS/WASM targets challenge React and TypeScript dominance.
First OWASP-backed security layer for ASI06 memory poisoning in agentic AI.
Pattern-based AI detection runs locally, but reliability remains fundamentally questionable.
CDP debugger primitives enable function hooks without monkeypatching the runtime.
56KB ML library beats TensorFlow.js for forecasting and smoothing, but niche problem.