Selector Forge is an innovative CSS and XPath selector generation and validation tool based on AI agents, open-sourced by Intuned, a company specializing in browser automation, on June 22, 2026. With the evolution of web applications, dynamic and complex DOM (Document Object Model) structures have become commonplace, making it nearly impossible to maintain stable web scraping or E2E (End-to-End) test code using traditional static methods. Selector Forge was created to overcome these limitations and is a next-generation solution that goes beyond the capabilities of conventional browser developer tools, which simply copy the physical hierarchical path of an element. It leverages an AI agent loop to analyze the DOM context of a web page from multiple angles and automatically generate semantic selectors. Designed with the WXT framework and React, this browser extension features a sophisticated architecture that communicates with a backend server in real-time when a user selects a target on the browser screen, allowing it to explore the most stable and robust set of candidate selectors. Traditional web automation tools have selector extraction features that are vulnerable and can be rendered useless with even minor changes to the HTML structure of a web page. The inefficiency of the entire scraping pipeline being halted by even slight changes, such as dynamically changing class names or modified layout arrangements, has been a major headache for developers. Just as an experienced GPS detects changes in road conditions in real-time and re-explores the optimal route, Selector Forge detects changes in the DOM structure of a web page in real-time and independently verifies and finds the optimal selector path that will never break. The most unique aspect of Selector Forge is that it does not immediately return the selector candidates suggested by the AI to the user but instead performs stress testing directly in the DOM environment of the currently active live page to cross-validate whether it can uniquely identify the target element. By providing only these high-confidence semantic selectors that pass this verification stage, it ensures resilience, meaning that it not only works once but also continues to function robustly even as the service is updated over the long term. In practice, this tool provides significant productivity gains for operators of large-scale web crawling pipelines, QA automation engineers, and developers of AI-based autonomous web agents. In particular, it supports a list mode that allows for the batch extraction of list-type data areas that appear repeatedly within a web page, not just individual elements, enabling complex unstructured data such as gene information tables or drug bio-activity lists to be accurately mapped with a single drag and click. Currently, Selector Forge offers convenient accessibility by allowing users to install it directly from the official extension store and guarantees 200 free AI selector generations per month. Furthermore, the developer, Intuned, is preparing a self-hostable backend open-source release to ensure data security and handle large requests in enterprise environments, giving it excellent scalability potential, allowing biotechnology research institutes or corporate research institutes that use closed network environments to deploy it directly on their infrastructure and use it as an independent data collection engine without worrying about external data leakage.

AI agent-based tool for automatically generating CSS selectors.