The latest articles on DEV Community by Anal Shaju (@as707). https://dev.to/as707 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%2F3589888%2F4dd4ea6b-12f4-4a21-92f3-39d19a0834e3.png https://dev.to/as707 en Anal Shaju Thu, 21 May 2026 05:19:02 +0000 https://dev.to/as707/i-built-a-cli-tool-that-converts-messy-webpages-into-clean-markdown-for-ai-tools-4964 https://dev.to/as707/i-built-a-cli-tool-that-converts-messy-webpages-into-clean-markdown-for-ai-tools-4964 <h2> The Problem </h2> <p>Every time I paste a webpage into Claude or ChatGPT, I waste my tokens on navigation bars, footers, ads, and junk.</p> <h2> What I Built </h2> <p>grabctx - a CLI tool that strips all the unwanted things and gives you only the main content as clean markdown.</p> <h2> How It Works </h2> <div class="highlight js-code-highlight"> <pre class="highlight shell"><code>npm <span class="nb">install</span> <span class="nt">-g</span> grabctx grabctx https://any-article.com <span class="nt">--copy</span> </code></pre> </div> <p>It fetches the page, extracts the main content using Mozilla's <br> Readability algorithm (same one Firefox uses for Reader View), <br> converts it to markdown, and shows you the token savings.</p> <h2> Results </h2> <p>Tested on a Wikipedia page:</p> <ul> <li>Before: 91,461 tokens</li> <li>After: 19,757 tokens </li> <li>Saved: 78%</li> </ul> <h2> Tech Stack </h2> <ul> <li>TypeScript</li> <li>Node.js</li> <li>@mozilla/readability</li> <li>linkedom</li> <li>turndown</li> <li>commander</li> <li>gpt-tokenizer</li> </ul> <h2> Links </h2> <ul> <li>npm: <a href="https://www.npmjs.com/package/grabctx" rel="noopener noreferrer">https://www.npmjs.com/package/grabctx</a> </li> <li>GitHub: <a href="https://github.com/AnalShaju/grabctx" rel="noopener noreferrer">https://github.com/AnalShaju/grabctx</a> </li> </ul> <p>Would love feedback from the community!</p> typescript cli opensource ai