The latest articles on Forem by Admir Šaheta (@admirsaheta). https://forem.com/admirsaheta https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F930511%2F2b094bd0-e177-4c8e-8edc-6755da15a7af.jpeg https://forem.com/admirsaheta en Admir Šaheta Sun, 30 Nov 2025 09:16:57 +0000 https://forem.com/admirsaheta/how-i-built-flyscrape-a-typescript-web-scraper-that-actually-works-on-modern-websites-4d9g https://forem.com/admirsaheta/how-i-built-flyscrape-a-typescript-web-scraper-that-actually-works-on-modern-websites-4d9g <p>If you’ve ever tried scraping modern websites, you know the pain. SPAs that never fully load, lazy images hiding content, bot protections blocking your every request… I was losing hours, sometimes days, just trying to get usable data.</p> <p>One late night, fueled by coffee and frustration, I thought: <strong>There has to be a better way.</strong> That’s how <strong>FlyScrape</strong> was born — a <strong>TypeScript Node.js package</strong> that makes scraping <strong>fast, reliable, and developer-friendly</strong>.</p> <p><strong>My Developer Nightmare</strong></p> <ul> <li>Multiple failed scrapers on SPAs </li> <li>Infinite scroll pages that never end </li> <li>Bot protections killing my requests </li> <li>Messy, unstructured HTML that no AI model could digest </li> </ul> <p>Sound familiar? I knew there had to be a smarter approach.</p> <p><strong>Enter FlyScrape</strong></p> <p>FlyScrape is <strong>modular, typed, and production-ready</strong>. It combines Playwright, stealth techniques, and AI-powered content cleaning to deliver <strong>structured, LLM-ready Markdown</strong> from any web page. </p> <p>I can now scrape <strong>dynamic content, handle infinite scrolls, and avoid bot detection</strong> — all with minimal code and zero frustration.</p> <p><strong>Why I Built It This Way</strong></p> <ul> <li> <strong>TypeScript-first</strong>: Strong types and autocomplete mean fewer bugs and faster development </li> <li> <strong>AI integration</strong>: Converts messy HTML to clean Markdown for RAG pipelines or content aggregation </li> <li> <strong>Extensible hooks</strong>: Customize every stage of the crawl </li> <li> <strong>Performance-minded</strong>: Caching, lazy load handling, and resource blocking keep things fast </li> </ul> <p><strong>How It Changed My Workflow</strong></p> <p>Before FlyScrape, I’d spend hours debugging failed crawls. Now I can: </p> <ul> <li>Spin up a crawler in minutes </li> <li>Extract content-ready Markdown in seconds </li> <li>Integrate with AI pipelines effortlessly </li> <li>Spend my time analyzing data, not wrestling websites </li> </ul> <p><strong>TL;DR</strong></p> <p>If you’re a <strong>developer tired of fighting modern websites</strong>, FlyScrape is here to save your sanity. Clean, modular, TypeScript-first, and AI-ready — built by a developer who’s been there. </p> <p>Check it out on GitHub: <strong>FlyScrape</strong> ( <a href="https://github.com/flyrank-bih/flyscrape" rel="noopener noreferrer">https://github.com/flyrank-bih/flyscrape</a> ) and star it ⭐ if you find it useful.</p> webdev ai programming javascript Admir Šaheta Thu, 22 Sep 2022 07:15:49 +0000 https://forem.com/admirsaheta/sheetster-custom-sheets-cjc https://forem.com/admirsaheta/sheetster-custom-sheets-cjc <p>The other day on a Swift discord I see people having a discussion how hard it is to set custom detents for sheets right now. And it inspired me to write a library that let's you automate and set custom detents for sheets, wether it'd be a bottom sheet like the one the maps use, or you just have a designer who places 30% 40% 70% height sheets into the UI Design.<br> If you find it useful, make sure to star it :)<br> <a href="https://github.com/admirsaheta/Sheetster">https://github.com/admirsaheta/Sheetster</a></p> swift github opensource ios