roamer.ai

Starting Page
Select Language
Navigation
Play Tour

Inspiration

Our inspiration came from Rudransh’s experience moving to Canada. Living in a French-speaking province, exploring his new neighborhood was difficult and isolating—signs provided little information in a language he understood. Exploration of new places can be one of the most powerful sources of growth, so we set out to augment those experiences.

Our goal: let anyone, anywhere, from any language background, seamlessly learn about the history and context of their surroundings.

What It Does

roamer.ai provides an audio stream of information and history about your immediate surroundings. Simply open the website, allow location sharing, and receive tour-guide-style descriptions—in any language.

Key Features

🌍 Omnilanguage selection 🎧 Seamless audio narration 📍 Geographically tailored information

The Build

Frontend:

React, Vite

Backend:

Flask, Python, Google Maps API, Google Gemini 2.5 API, Eleven Labs API, Gunicorn

Key Partners:

Eleven Labs – AI-powered text-to-speech
Google Gemini 2.5 – LLM-based historical + contextual descriptions

Challenges We Faced

Adding accessible features, creating captions and multiple language options within the existing design led to a difficult retrofit.
Integrating frontend and backend smoothly, the constant need for updated coordinates provided a unique challenge.
Handling query latency from API requests, due to the complexity of the prompts asked we had to navigate how a user would be affected by delay.
Preventing repeated information when standing still, we had to ensure the system was not repeating itself if users did not move location.