LandingAgent: A Reference-Annotated Dataset and Agentic Generation Framework for Landing Pages

LandingAgent is a new framework for generating landing pages that are tailored to a specific target. It uses a reference‑annotated dataset called LandingBench, which abstracts real landing pages into structured elements such as section sequences, layout patterns, tone descriptors, visual emphasis, and CTA structure. The agentic framework operates in three phases—profiling the target, building a reference‑guided wireframe, and refining the page through critique—resulting in pages that are more faithful to the target, concise, readable, aesthetically pleasing, and structurally diverse compared to direct prompting.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Computation and Language.

• llms
• agents