LLM-Assisted Web Development: A Systematic Review of Capabilities, Tools, and Open Challenges Across the Web Technology Stack

Large language models have moved from writing isolated code snippets to taking a part in full web development workflows. This review maps that involvement across the languages and frameworks in current use: markup and styling (HTML, CSS, Tailwind CSS), core scripting languages (JavaScript, TypeScript), front-end frameworks (React, Vue.js, Next.js), back-end and data layers (Node.js REST APIs, GraphQL, SQL databases), and the testing layer (end-to-end and GUI-level test generation). Drawing on peer-reviewed studies and benchmark papers that specifically evaluate one of these layers, the review finds substantial, unevenly distributed research coverage: React and JavaScript dominate the literature, TypeScript and Vue.js have smaller but growing pockets of research, and several widely used pieces of the stack, including Svelte, Angular, and NoSQL databases such as MongoDB, currently have little or no peer-reviewed coverage of their own. The review also brings together the available evidence on productivity, security, and developer trust, and closes with a discussion of open problems specific to each layer of the stack.