Local-ish

In Bali, I suddenly felt rich. Five dollars became around eighty thousand Indonesian rupiah. I had all these banknotes in my hand without understanding what things were actually worth. A price could sound cheap in dollars and still be much higher than what locals usually paid. After being overcharged, I realised a currency converter couldn’t answer my real question: “Is this a fair price here?”

Travelling with friends added another challenge: remembering who paid and who owed whom, without making money an awkward part of the trip.

These experiences inspired Local-ish—a way to understand unfamiliar prices, track spending, and remember where money goes. Our one goal is removing the friction travellers face when making everyday financial decisions in unfamiliar places.

We built a web app with a TypeScript and Express backend, dividing our work between interface, data, and Voice & AI. Gemini interprets typed requests, while ElevenLabs handles transcription and spoken phrases. Users can review an expense, compare it with available reference prices, and save it to their money balance.

We learned that understanding a price requires context: “two teas for sixty” is different from “sixty each.” We also learned to use AI for interpretation and reference data for comparisons, rather than asking AI to invent a fair price. Voice input, API quotas, and connecting our separate parts were challenges, which we addressed through transcript review, fewer repeated requests, shared contracts, and automated tests.

This hackathon taught us that building a useful AI application isn't simply about integrating a chatbot. It's about making different technologies work together to solve a real problem. We gained practical experience with APIs, frontend and backend development, databases, AI integration, testing and team collaboration. We're especially proud of connecting our backend to TigerData, storing structured price references and combining price checking, budgeting and bargaining assistance within an interactive prototype.

We want Local-ish to become a smarter and more personalised travel-spending companion. Our next steps include: Budget-aware recommendations: Connecting the price checker with the traveller's remaining budget so it can evaluate both price fairness and affordability. Photo-based price checking: Allowing travellers to photograph menus, taxi tariffs and price signs instead of typing everything manually. Connected group accounts: Enabling friends to link their accounts and manage shared expenses more seamlessly. More reliable price information: Expanding our database to additional destinations and improving the freshness and verification of our reference prices. Our long-term vision is to make local price knowledge accessible to travellers everywhere, helping them make financial decisions with greater confidence.