VON Travel Assistant

HarmonyOS Junior Developer, Yingdao Junior RPA Engineer, YashanDB V23.2 Certified Admin, Kingbase Certified Specialist. Led 10 provincial team competitions, 2 individual competitions with provincial honors. In 2025, led team to win Henan Province First Prize in 'Challenge Cup', OpenAgents Community First Prize, Hacker Pioneer Award.

A multi-agent travel assistant system developed based on OpenAgents framework, providing intelligent travel services through multi-agent collaboration.

Planning 4 Agents, building a collaborative multi-agent travel assistant system on OpenAgents framework. Agent Roles: travel_router (Receptionist) - welcomes users, introduces features, handles general inquiries (like 'hello', 'what can you do'). weather_agent (Weather Assistant) - queries weather, provides clothing advice (responds to 'weather', 'temperature', 'rain' keywords). spot_agent (Attraction Assistant) - recommends popular attractions, introduces guides (responds to 'attraction', 'recommend', 'West Lake' keywords). plan_agent (Itinerary Assistant) - plans multi-day trips, arranges routes (responds to '3-day itinerary', 'how to play', 'route' keywords). Collaboration: All Agents join same OpenAgents Network via gRPC on local port 8700. User messages broadcast to all Agents, each self-determines whether to respond based on keyword rules. Decentralized listening mode avoids single-point bottlenecks, improves response flexibility.

① travel_router Agent (Receptionist): Responds to user greetings ('hello', 'hi'), introduces system features, handles non-professional travel inquiries, proactively avoids messages with 'weather', 'attraction', 'itinerary' keywords. ② weather_agent (Weather Assistant): Parses weather queries ('Is Beijing cold today?', 'Need umbrella?'), provides current/forecast weather info and clothing advice based on LLM knowledge. ③ spot_agent (Attraction Assistant): Identifies city and attraction requests ('Recommend fun places in Hangzhou', 'Must-visit attractions in Xi'an?'), returns structured attraction lists with name, location, features, recommendation level. ④ plan_agent (Itinerary Assistant): Understands multi-day trip needs ('Plan 3-day Shanghai trip', 'How to play 2 days in Chengdu?'), generates detailed itinerary with daily schedule, attraction order and practical tips based on geographic proximity and classic priorities. ⑤ Four Agents collaborate through OpenAgents Network: Share same message channel, user input broadcast to all Agents, each independently decides whether to respond based on YAML keyword rules, achieving smart distribution without central scheduling.