AI Travel Planner

Problem

Travel planning often starts with a thin prompt like Tokyo 3 days, then still requires follow-up questions about dates, budget, and interests. Outputs can also feel scattered, hard to paste into Notion or Feishu, and missing a shareable visual summary. insight-travel-guide addresses this by treating destination and trip length as the minimum input, then producing a Markdown itinerary and optional HTML travel card.

How It Works

• Structured itinerary: Covers overview, daily plans, lodging, food, transit, and budget. Daily entries are split into morning, lunch, afternoon, dinner, and evening, with suggested duration and estimated cost.
• Requirement filling: Normalizes phrases such as three days, budget 10000, and cheap. Only destination and days are required; missing critical fields are confirmed before generation.
• Generation and validation: Uses built-in model capability without external APIs. It checks section completeness, day count consistency, and budget totals, then retries or falls back to a simplified itinerary on failure.
• Shareable card: With user confirmation, generates an HTML card with city header image, attraction photos, daily thumbnails, and budget summary, including image fallback handling.
• Iteration: Supports replacing a single day, compressing budget, adding restaurants, adding nearby attractions, or changing card colors and images.

Boundaries

Best for 1-14 day city or regional trip planning. It does not perform real-time booking, live navigation, real-time pricing, calendar sync, or visa services. Hotel and restaurant prices are AI-estimated reference values. When online, it can add weather, festivals, visa notes, and exchange-rate hints; offline it falls back to built-in knowledge and marks non-real-time information.