Inspiration
The inspiration for VoyagerAI 2.0 came from the frustration of planning travel across dozens of open browser tabs and busy personal calendars. We wanted to build a proactive agent that doesn't just guess your travel dates, but actually knows when you are free by checking your real schedule.
What it does
VoyagerAI 2.0 is a split-screen AI travel agent. On the left, an animated avatar speaks to you using a synthetic voice. On the right, a smart chatbot scans your Google Calendar, searches live flight and hotel prices, and builds a complete day-by-day itinerary. Once your trip is finalized, it provides a dashboard to download a professional PDF, email the plan, or sync it directly to your calendar.
How we built it
We used the newest Next.js 16 and React 19 frameworks with Tailwind CSS v4 for a premium look. The intelligence is powered by Groq and Gemini AI models. We used SerpApi to get live flight and hotel data, Google Cloud for secure calendar access, and the jsPDF library to generate the final travel documents.
Challenges we ran into
One of our biggest challenges was the AI response getting cut off halfway through because the itineraries were so detailed. we fixed this by increasing the allowable token limit and teaching the AI to be brief in the chat while putting all the detail into the JSON code. We also had to build complex logic to automatically find the nearest airport to the user's current city.
Accomplishments that we're proud of
We are especially proud of the "Trip Finalized" dashboard. It feels like a state-of-the-art tool that smoothly handles the transition from a text chat to a functional travel document. We are also proud that the entire app is built with strict TypeScript and is ready for full production deployment.
What we learned
We learned that an AI is only useful if it is grounded in real data. By building strict logic for budget estimation—including flights, hotels, and daily spend—we ensured that the AI never suggests a trip that the user cannot afford. We also learned how to orchestrate multiple powerful APIs into a single, seamless user experience.
What's next for Voyager AI 2.0
Next, we want to add real checkout links so users can book their flights and hotels directly through the app. We also want to implement a group mode where multiple friends can link their calendars and plan a trip together in one shared chat window.
Built With
- and-tailwind-css-v4
- elevenlabs-(tts)
- gcp
- gemini
- google-calendar-api
- googlecloudplatform
- groq-ai-(llama-models)
- next.js
- nextauth.js
- react
- serpapi-(flights/hotels)
- tailwindcss
- typescript
- utilizing-google-cloud-platform
Log in or sign up for Devpost to join the conversation.