TripCraft: A Benchmark for Spatio-Temporally Fine Grained Travel Planning

TL;DR

TripCraft tackles the gap between real world travel constraints and LLM generated itineraries by constructing a high fidelity, constraint aware travel planning benchmark. It combines transit schedules, events, attractions, and persona factors for 1000 queries across 140 U.S. cities, and introduces five continuous evaluation metrics to assess meal timing, attraction durations, travel efficiency, sequencing, and persona alignment. The dataset includes gold standard plans annotated by 25 humans, enabling fine-grained assessment beyond binary feasibility. Experiments with GPT-4o show that parameter informed prompting improves certain objective metrics such as meal scheduling, while also revealing trade-offs with constraint adherence, establishing TripCraft as a practical benchmark for advancing AI driven itinerary generation.

Abstract

Recent advancements in probing Large Language Models (LLMs) have explored their latent potential as personalized travel planning agents, yet existing benchmarks remain limited in real world applicability. Existing datasets, such as TravelPlanner and TravelPlanner+, suffer from semi synthetic data reliance, spatial inconsistencies, and a lack of key travel constraints, making them inadequate for practical itinerary generation. To address these gaps, we introduce TripCraft, a spatiotemporally coherent travel planning dataset that integrates real world constraints, including public transit schedules, event availability, diverse attraction categories, and user personas for enhanced personalization. To evaluate LLM generated plans beyond existing binary validation methods, we propose five continuous evaluation metrics, namely Temporal Meal Score, Temporal Attraction Score, Spatial Score, Ordering Score, and Persona Score which assess itinerary quality across multiple dimensions. Our parameter informed setting significantly enhances meal scheduling, improving the Temporal Meal Score from 61% to 80% in a 7 day scenario. TripCraft establishes a new benchmark for LLM driven personalized travel planning, offering a more realistic, constraint aware framework for itinerary generation. Dataset and Codebase will be made publicly available upon acceptance.

Figures (6)

• Figure 1: TripCraft overview: Continuous evaluation metrics along-with fine-grained itinerary details.
• Figure 2: Visualization of the Construction Pipeline.
• Figure 3: Existing constraint-based metrics rate these plans equally, but, our metrics provide a continuous score, capturing temporal, spatial, sequential, and personal differences.
• Figure 4: Age Distribution of our graduate student annotators.
• Figure 5: Gender Distribution of our graduate student annotators.