Story

I'm a dev who got tired of wasting entire evenings planning trips. After the 10th time finding myself with 50 tabs open at 2 AM reading contradictory blog posts, I built Voyaige.

What it does: Enter a city + your travel style (budget backpacker, luxury, foodie, etc.) → get a personalized PDF guide in ~15 minutes. Not generic "Top 10" lists—actual itineraries with opening hours, transport tips, and recommendations that match YOUR preferences.

Why Perplexity over GPT/Claude: Perplexity's Deep Research API was the game-changer. It actually cites real sources and pulls fresh data. GPT-4 kept giving me outdated restaurant recommendations from 2021. Perplexity searches, synthesizes, and cites—perfect for travel where accuracy matters.

  Tech stack:
  - Laravel backend with queue workers for long-running generation
  - Perplexity Deep Research API for research + synthesis
  - Custom PDF generation (tried DOMPDF, settled on Browsershot + headless Chrome)
  - Polar for payments (Stripe rejected me 3 times as "travel content")

  Hardest technical challenges:
  1. API response quality: Prompt engineering to get consistent structure across different cities/personas
  2. PDF layout: Making 25-page guides that actually look good and are readable on mobile
  3. Queue management: Handling generation failures gracefully + retry logic
  4. Payment processors: Finding one that accepts "AI-generated content" businesses

  Questions for HN:
  1. Would you pay $13 to skip 3-5 hours of research?
  2. What would justify higher pricing? 
  3. How do you feel about AI-generated travel advice vs. human travel bloggers?

  Link: https://voyaige.io

  Happy to discuss or answer any questions!