TaskMaster AI

AI-powered task management app offering geolocation, advanced inference, insights, automation, and location-based recommendations.

Comment

Inspiration

Task management is overwhelming. We've personally tried too many systems that don't work for us -- especially having ADHD, it's too easy to overschedule, hyperfocus or to just be plain unhealthy about how we go about our weeks (let alone forgetting about tasks entirely). We wanted to build out a method that would work for us, along with other people who have the same issues.

Our goal was therefore to create a task manager that can organize tasks based on minimal input, show them as markers on a map based on whether a street address was given, and provide intelligent insights based on user behaviour and location. We wanted to make the interface as low-barrier as possible to maximize adoption.

What it does

TaskMaster AI helps users manage tasks efficiently by leveraging AI (Perplexity, AWS Bedrock), MongoDB Atlas, Cloudflare, and map APIs to:

  • Automate task input through natural language processing, allowing users to enter tasks in natural English. AI infers the relevant information by segmenting their input out, no hard-coding required!
  • Game-ify task completion by showing tasks as complete-able markers on a map.
  • Provide personalized insights based on task patterns and location.
  • Offer recommendations for task balancing and time management.

How we built it

We stitched together many backend and frontend technologies:

  • Backend: Built with Node.js, Express, and MongoDB for persistent task storage, user management, and API integration with AWS Bedrock and Perplexity Pro.
  • Frontend: Developed using React with integration of Mapbox's API for displaying and managing location-based tasks.
  • AI Integration: AWS Bedrock was used to analyze task patterns and provide insights, while Perplexity Pro helped automate task entry through segmenting natural language input.

Challenges we ran into

  • Google Cloud issues: Google Maps declined our billing info, likely because we were trying to add it from a different city than what Google would expect. We had to switch to and pick up Mapbox last-minute, but we made it work well!
  • Integrating multiple AI services: We were working with many of these APIs for the first time, and we had to balance the outputs from AWS Bedrock and Perplexity to ensure smooth task automation and insights generation.
  • Securing data: Managing API keys and securing user data in MongoDB Atlas while ensuring smooth cloud service access.

Accomplishments that we're proud of

  • Successfully integrating AWS Bedrock and Perplexity Pro to enhance user experience with AI-driven insights and automation.
  • Implementing real-time location-based task suggestions using Google Maps API.
  • Building a scalable architecture that handles user data efficiently while providing personalized task recommendations.

What we learned

We learned how to effectively:

  • Leverage AWS Bedrock for analyzing task patterns and generating insights.
  • Utilize Perplexity Pro to process natural language input and convert it into structured task data.
  • Integrate location-based services like Google Maps API to provide more personalized, context-aware task management.

What's next for TaskMaster AI

  • Enhanced insights: Expanding AI insights to cover more areas of time management, task prioritization and historical analysis. Time limitations didn't allow us to pursue Bedrock-related options as much as we'd hope to!
  • Improved automation: Further refining natural language processing to make task entry even more seamless.
  • Mobile integration: Developing a mobile version of the app to allow users to manage tasks on the go with location-based alerts.

Built With

Share this project:

Updates