Interwebcoder improves interactive webpage coding with fast test feedback
InterWebCoder: Learning Interactive Webpage Generation from Executable Tests via Reinforcement Learning
Software Engineering
Summary
Making interactive webpages that look right and respond correctly can be really tricky, especially when using AI models to generate the code. The authors created InterWebCoder, a system that learns from real user interactions and uses automated tests to check if the webpage works and looks correct. Instead of slow and unreliable methods, it quickly finds and fixes errors based on test results, improving both accuracy and training speed. Their approach outperforms larger models while using less computing time.
What this means in practice
- •For web developers: Generate interactive webpage code that reliably meets visual and functional requirements by learning from executable user interaction tests.
- •For software testing teams: Speed up testing and debugging of interactive web applications using execution-verified synthetic web code to isolate and fix errors efficiently.
Authors
Jingyu Xiao, Dasen Dai, Yuxuan Wan, Jen-tse Huang, Yintong Huo, Hongyu Zhang, Michael R. Lyu
Abstract
Interactive webpages underpin modern applications, but reproducing both their visual appearance and multi-step behavior remains challenging for multimodal large language models (MLLMs). Scalable training data must capture user actions, state transitions, and expected outcomes, while real-world webpage code is often entangled with irrelevant scripts and external dependencies. Moreover, existing GUI-agent-based verification provides slow and potentially unstable functional feedback. To address these challenges, we propose InterWebCoder, an efficient framework for interactive webpage code generation. InterWebCoder introduces a Test-Specification-Guided and Execution-Verified Data Synthesis pipeline that records real-world interaction trajectories and synthesizes self-contained webpage code with executable test scripts. A failureattributed repair mechanism uses runtime evidence to locate and iteratively fix errors in either the webpage code or test script. We then combine test-ID-weighted supervised fine-tuning with group-relative policy optimization, directly optimizing functional correctness and visual fidelity using test pass rates and state-level visual similarity. This replaces costly GUI-agent verification with fast, reproducible executable feedback. Our pipeline produces 1,736 verified samples from 2,058 tasks, achieving an 84.35% synthesis success rate. On IWR-Bench and Vision2Web, InterWebCoder achieves final scores of 48.54 and 48.42 with a 9B backbone, outperforming 30B-level models and Gemini-3.7-Flash while remaining competitive with Claude-Sonnet-5. It reduces reward-computation latency by over 97%, eliminates external MLLM verification costs, and accelerates end-to-end training by up to 1.4x.