Dropby | Devpost

Try it now at https://dropby-htn24.netlify.app

video demo in gif

Inspiration:

The idea for DropBy was born out of a simple, everyday frustration. One day, I was craving Tim Hortons coffee, but I didn't want to waste time if the line at the SLC location was too long. I thought, "There must be someone there right now who could tell me about the line, if only I had a way to ask them." This made me realize that there's a wealth of real-time, hyper-local information all around us, locked away in the minds of people nearby. DropBy aims to unlock this potential, connecting people who need immediate, local information with those who have it, creating a network of real-time, community-driven intelligence.

What it does

DropBy connects users with nearby helpers who can provide instant assistance or information. Whether you need a quick errand run, real-time local information, or help with a task, you can reply on DropBy's users' immediate help.

How we built it

We built our frontend with React + Typescript. For backend, Convex's realtime native API proved to be invaluable, allowing us to create a responsive and scalable system, which is especially useful when handling instant updates of real time information. Additionally, we used Cohere and OpenAI's APIs for AI features such giving a task a short and concise title and content moderation.

Challenges we ran into

Ensuring user safety and trust was our biggest challenge. We want to make sure the tasks on DropBy is safe and appropriate. To solve this challenge, we leveraged Cohere and OpenAI's moderation APIs to implement multiple safeguard against inappropriate user input and content generation. We also implemented a downvote feature so the community can remove any inappropriate content by downvoting.

Accomplishments that we're proud of

We've built a responsive platform using Convex's realtime capabilities and created a user-friendly interface. Our AI-powered system automatically generates concise task titles, categories, and time estimates from user descriptions, streamlining the experience and ensuring clarity across all tasks. We also successfully developed a user-friendly interface that simplifies complex interactions, making it easy for users to request and offer help.

What we learned

Diving into real-time data was totally new territory for us, and thanks to Convex, it was a breeze. We also got a real sense of how important it is to keep things clean and appropriate when you're inviting the public to join in. Plus, we tried out using more structured responses from AI language models for the first time, and it's a game-changer—they're way more reliable now.

What's next for DropBy

Moving forward, we plan to roll out user authentication for added security. We're also designing a points system to motivate and reward users for their contributions. An exciting feature in development is an AR companion that displays quest locations in real life and on a mini-map, similar to open-world RPG games. While we couldn't complete this due to time constraints and limited Unity experience, it remains a key goal for our next iteration.