OpenAI Build Week MVP

Plan, monitor, and replan one trip workspace.

A direct second-build inspired by TripStar: multi-agent planning, map-first results, budget intelligence, weather-aware replanning, knowledge graph, and contextual AI chat.

Open workspace

123

Budget fit94%

Weather stateAll clear

Memory3 prefs

01Requirement agent

02Discovery agent

03Weather agent

04Stay agent

05Route agent

06Graph builder

ReadyRequirement agent

Asks missing questions and reads preference memory.

Plan code AITC-0812

Sydney to Bali, 5 days for 1 traveler

The copilot consolidates scattered planning tasks into one surface: visa checks, route clusters, weather risk, budget split, and preference memory.

Route qualityLow backtracking

Visa riskArrival checklist

Live signalAll clear

Live signals and replan

All clear

Keep the original route and timing buffers.