Xiaoming's Road Trips

Users often know only a destination like “Dali road trip” or an EV route like “Xiangri 318”, but trip planning is still fragmented across attractions, restaurants, souvenirs, charging, and lodging. Xiaoming's Road Trips turns that into an executable route with verifiable locations, comparable prices, and charging checkpoints.

How It Works

• Intent detection: distinguishes destination queries, food questions, souvenir questions, route comparisons, itinerary edits, and EV-specific charging needs.
• Retrieval and synthesis: prioritizes search(source="kb") for the built-in road-trip knowledge base, then uses search(source="web") for Xiaohongshu notes, Amap street-level food rankings, geographical-indication products, charging stations, and online store prices.
• Structured output: emits ten required modules: overview, attractions, classic food with old shops, snacks, souvenirs with price comparison, food streets, gift shops, EV charging, an N-day itinerary, and practical tips.
• Key constraints: restaurant addresses should be precise to street numbers, souvenirs should compare onsite and online prices, attractions should include tier and popularity, and itineraries should include lodging, meals, and charging points.

Boundaries

This skill is best for pre-trip planning and information consolidation, especially for fuel and EV road trips. It relies on a knowledge base and search results, so shop status, charging power, local prices, and traffic rules may change and should be rechecked before travel. Real-time navigation, ticket booking, or vehicle telemetry still require maps, ticketing platforms, or vehicle systems.