FullStack-Agent: Enhancing Agentic Full-Stack Web Coding via Development-Oriented Testing and Repository Back-Translation

TL;DR

FullStack-Agent addresses the real-world gap in producing production-grade full-stack web applications by coupling a planning-driven multi-agent development framework with a data-centric backbone learning loop and a comprehensive full-stack benchmark. The system demonstrates substantial gains over prior agentic baselines across frontend, backend, and database tasks, and shows that iterative self-improvement through repository back-translation and augmentation yields meaningful performance boosts for even smaller models. Specialized debugging tools and sandboxed execution enable efficient error localization and robust development workflows. The work also provides a scalable benchmark (FullStack-Bench) to evaluate end-to-end functionality, facilitating future research in agentic software engineering and highlighting important societal considerations for automated code generation.

Abstract

Assisting non-expert users to develop complex interactive websites has become a popular task for LLM-powered code agents. However, existing code agents tend to only generate frontend web pages, masking the lack of real full-stack data processing and storage with fancy visual effects. Notably, constructing production-level full-stack web applications is far more challenging than only generating frontend web pages, demanding careful control of data flow, comprehensive understanding of constantly updating packages and dependencies, and accurate localization of obscure bugs in the codebase. To address these difficulties, we introduce FullStack-Agent, a unified agent system for full-stack agentic coding that consists of three parts: (1) FullStack-Dev, a multi-agent framework with strong planning, code editing, codebase navigation, and bug localization abilities. (2) FullStack-Learn, an innovative data-scaling and self-improving method that back-translates crawled and synthesized website repositories to improve the backbone LLM of FullStack-Dev. (3) FullStack-Bench, a comprehensive benchmark that systematically tests the frontend, backend and database functionalities of the generated website. Our FullStack-Dev outperforms the previous state-of-the-art method by 8.7%, 38.2%, and 15.9% on the frontend, backend, and database test cases respectively. Additionally, FullStack-Learn raises the performance of a 30B model by 9.7%, 9.5%, and 2.8% on the three sets of test cases through self-improvement, demonstrating the effectiveness of our approach. The code is released at https://github.com/mnluzimu/FullStack-Agent.

Figures (22)

• Figure 1: The FullStack-Agent system. It combines a multi-agent development framework equipped with efficient coding and debugging tools (FullStack-Dev), an iterative self-improvement method that enhances LLMs through repository augmentation and back-translation (FullStack-Learn), and a comprehensive benchmark evaluating frontend, backend, and database functionalities (FullStack-Bench).
• Figure 2: Prompt used in baseline testing for generating a full-stack website.
• Figure 3: The error composition of the frontend, backend, and database tests.
• Figure 4: The manual annotation interface for frontend tests, showing the GUI-agent trajectory that can be played as a video, and the corresponding database interaction logs.
• Figure 5: The manual annotation interface for backend tests, showing backend API testing trajectory, and the corresponding database interaction logs.