ARtifact

The world is your museum. With ARtifact, every building could tell you its story.

Comment

Inspiration

ARtifact was inspired by our laziness, always having to search up landmarks and their info when travelling. We wanted an app that could act as a personal tour guide: just point your camera, and the app tells you everything.

What it does

ARtifact is a multi-tool for travellers. It has three core features:

  1. AI Lens: Users can point their camera at a landmark, and the app uses an AI model to identify it and provide information and fun facts, acting as an on-demand tour guide.
  2. Planner: Using the device's GPS, the Planner feature fetches a list of nearby points of interest.
  3. Translator: A simple but effective translation tool for extra convenience, ensuring users are never without a way to communicate.

How we built it

The app was built using React Native and Expo. The development process was feature-focused:

  1. AI Lens: The primary feature uses expo-camera to capture an image. This image, along with GPS data from expo-location, is sent to an AI model, alongside a special prompt. For now, we use the free grok-4-fast API, and fall back on the OpenRouter API.
  2. Planner: This feature uses the Wikipedia Geosearch API. It takes the user's current coordinates and fetches a list of nearby points of interest to help them discover their surroundings.
  3. Translator: This feature defaults to the Google Translate API and falls back to LibreTranslate instances if the primary service fails, ensuring functionality.

Challenges we ran into

  • API Latency: The AI image analysis was slow. We mitigated this by implementing a timeout utility to prevent the app from freezing on slow requests.
  • Time Constraints: Our team joined the hackathon relatively late and had to balance between school work as well as other competitions.
  • Demo Video We did not have aligned free time, so we couldn't come together to film a demo video. Furthermore, none of us knew how to use video editing software, so we had to learn everything from scratch.

Conclusion

Coding using AI was certainly an experience. It definitely sped up the beginning process a lot, but we quickly got rate-limited, perhaps due to our inexperience with efficient prompting. I could definitely see this being the coding norm in the future, though.

Learning how to use the OpenShot video editor was also quite something. That took up an embarrassingly large amount of time, but we pulled through.

Overall, it was a fun project.

Built With

Share this project:

Updates