Yesterday's Top Launches: 1 Tools from June 24, 2026
Selector Forge is a new browser extension that uses AI to generate resilient CSS and XPath selectors for web scraping and testing, reducing the frustration of broken selectors.
Yesterday's single standout in the world of new developer tools wasn't a massive framework or a new programming language. Instead, it tackled one of those persistent, everyday frustrations that can grind a productive afternoon to a halt: the search for a reliable CSS or XPath selector. If you've ever spent twenty minutes trying to figure out why your web scraper or UI test broke overnight because a div got a new class, you'll understand the immediate appeal of this launch.
Selector Forge
This new browser extension, launching into what's become a surprisingly crowded space for selector tools, makes a compelling promise: to use AI not for gimmicks, but to actually understand what makes a selector resilient. The core problem it targets is the brittleness baked into most automated selector generation. Your browser's DevTools, or even dedicated extensions, typically spit out a path that's a precise but fragile snapshot of the DOM at that exact moment—a chain of div > div:nth-child(3) > span.my-class. It works until the frontend team ships a design tweak, and then everything falls apart. The maintenance burden this creates for automation scripts and scrapers is immense, turning what should be a set-and-forget task into a recurring chore.
Selector Forge approaches this differently. When you highlight an element on a page, its AI doesn't just map the shortest path. It analyzes the surrounding structure and semantic clues, actively looking for stable attributes to latch onto. The goal is to find identifiers that are meaningful, like an aria-label, a data-testid put there by engineers, or even consistent text content, rather than ephemeral class names generated by a CSS framework. It's a shift from "find this exact element now" to "find elements like this in a way that will still make sense later."
A feature that caught my attention is its handling of element arrays. Needing to select all items in a list or all rows in a table is a common task, yet generating a single clean selector for the whole set is often a manual, trial-and-error process. The extension can analyze multiple selected elements, identify their common structural or semantic pattern, and output one selector that targets the entire group. For anyone building data extraction pipelines, this alone could shave off a significant amount of setup time.
Practically, it works as you'd hope a browser extension should. You install it, click its icon, and interact with it directly on the live page you're working on. You get a panel where you can select elements and see the AI's suggested selectors in real time, complete with a count of how many elements on the page that selector currently matches. This immediate feedback is crucial—it lets you judge the specificity and accuracy of its suggestions before you copy anything into your code.
The honest observation here is that the success of this tool will live or die by the quality of its AI model. The concept is solid, even necessary, but if its suggestions are only marginally better than a careful human using DevTools, it becomes a nice-to-have rather than an essential. The promise of "semantic" selection is the key differentiator; we'll need to see how well it performs across the tangled, modern web of React components and dynamic content to judge its true impact.
Who stands to benefit the most? If you're doing any substantial web scraping, especially from sites that update frequently, this could be a game-saver for maintenance. For QA engineers and developers writing end-to-end tests with Selenium, Playwright, or Cypress, generating robust selectors is directly linked to reducing test flakiness. A test suite that doesn't break because a CSS class changed from btn-primary to btn--primary is a more trustworthy and cheaper-to-maintain test suite.
On the practical side, Selector Forge is launching with a generous free tier—200 selector generations per month—which is plenty for individual developers or small projects to test it thoroughly. For teams with heavier usage, paid plans offer unlimited access. Notably, the project is open-source, with the code available on GitHub, which is a strong signal of transparency and a welcome approach for a tool that needs to earn developer trust. Future integrations like a CLI or support for Model Context Protocol (MCP) could make it a seamless part of more automated workflows.
In essence, Selector Forge isn't just another utility; it's a focused attempt to apply intelligence to a defined, painful problem in the web automation stack. It won't write your code for you, but it might just stop your code from breaking so often.
Quick link from yesterday: