Integrate Bright Data's search and scraping APIs with Convex for reactive caching of search results and page content with configurable TTLs.

🚀 Enhance your HTTP requests with undici-extra, offering a familiar, elegant API while leveraging the performance of Undici.

Real-time AI chess coach with streaming analysis.

Streaming projection engine — extract fields at multi-gigabit speeds with O(1) memory

Frontend for FrameForge built with React 18, TypeScript, and Vite. It provides a workshop-style interface for generating or correcting React components, visualizing results live, reviewing code, navigating between design sessions, and browsing a version history of every code iteration the AI produced.

Unofficial real-time chess client -- Lichess.org API, OAuth 2.0 PKCE, live NDJSON streaming, type-safe generated API client.

Next.js web app for converting public web pages into clean Markdown with live progress, preview, copy, and download.

Deep AJAX interception for XHR & Fetch with unified request/response hooks and stream support (SSE/NDJSON).

VS Code extension to view and edit NDJSON/JSONL files in a spreadsheet-like grid with inline, docked, and flat subtable expansion

Parse Claude Code's --output-format stream-json into fully typed TypeScript events. Zero dependencies.

Convex component that wraps Bright Data's Datasets API. Trigger async collections, receive results via webhook, and subscribe to structured records in real time via useQuery.

A state-layout mapping runtime where the server owns state and the client owns DOM projection. MPA pages with NDJSON streaming updates.

A fight club for LLMs. 🤫 Live demo → https://chatwar.ai

Ultralight log sidecar for Node.js that turns NDJSON into instant queries and actionable webhook alerts.

ANEL: The POSIX for AI agents — structured execution layer with Unix pipes, NDJSON streams, Wasm sandboxes, and self-healing errors. 4-language reference implementation.

🌊 Elegant HTTP client for Node.js based on Undici with extra features

Efficiently render 50,000 streaming logs at 60fps and time to first byte under 5ms.

A serverless-friendly scraper API that accepts a list of URLs, runs Playwright against each one, and streams the results back as newline-delimited JSON (NDJSON). The request stays synchronous, but you get per-URL feedback as soon as each scrape completes

🚧 HTTP client based on the Fetch API with timeout, retries, streaming, and downloads for browser and Node.js.

Accept and automatically parse NDJSON stream requests in NestJS with Express!