Orbital AI

Meet Orbital AI. Your personal AI assistant that orbits around you. It can automate your daily workflow based on connections from the apps you use every day, from your Gmail inbox receiving work assignments, to meeting times scheduled based on your Slack mentions. All you need to do is click approve to add the task to your to-do list, letting you focus on the important things in your life. Keep your workflow on autopilot with Orbital’s use of agentic AI technologies, intelligent inference courtesy of Gemini 2.5 Flash Lite, and human-in-the-loop (HITL) design.

Inspiration

When considering what problems we could tackle with our current advancements in AI-agentic automation, we focused on what contributes the most to productivity loss among many people, particularly students. We looked into the idea of cognitive friction, repetitive tasks that would waste the mental energy of an individual. We realized that many people have to jump between different digital services to complete a single task, and imagined what it would be like to have one singular platform that managed it all. Better yet, a platform that could automate those tasks for you with the ease of pushing a button.

What it does

Orbital AI was created to turn our inspiration into reality, developing a platform where you are able to automate tasks on services like Gmail and Slack based on automated triggers. We offer two ways to automate these tasks: a chat mode that resembles a traditional AI chatbot and a task view, which shows real-time suggestions based on automated triggers. Both allow you automate your digital life across multiple platforms from a single service with your entire personal context in mind.

How we built it

This app is built using the Next.js 14 framework, completely foreign to all of us on the team. The frontend of this ambitious project is styled with Tailwind CSS and Framer Motion for the animations. Underneath all this is an AI layer where we utilize Composio, Gemini API, and Vercel AI SDK. This combination allows us to create AI agents capable of performing tasks such as retrieving information from Slack and deciding how it should deal with that information. Gemini API is used here as an inference model to help determine if the provided information is a task or normal chat. We also use Redis as a short-term memory to hold the tasks still waiting for approval from the user.

Challenges we ran into

A major challenge we ran into was the integration of Composio with Notion. The documentation has a lot of discrepancies between versions since some of the functions were out of date. OAuth was a very big obstacle since it prevented many integrations from connecting together. With the limited time, we decided it was a better use of time to just create our own simple calendar.

Accomplishments that we’re proud of:

Some accomplishments that we’re proud of were creating our own AI-agent that has different tools and being able to choose what tools it should use to accomplish the task. The scale of this project is also something we are proud of, since many of our projects didn’t quite have the scale that Orbital has. The technologies are all foreign to us, and the fact that we were able to manage a working product is something we will always be proud of.

What we learned

We learned Next.js, Redis, using OAuth, and Gemini API. Next.js was completely foreign to us, and the fact that we were able to utilize it for our needs was a massive learning moment. OAuth was also a major learning lesson, as many of our projects previously didn’t include it, making it something very valuable in our learning experience with Orbital. Additionally, by using these industry-standard technologies, we learned multiple skills of problem-solving (quickly!), team-building, and programming all within the 24-hour window.

What’s next for Orbital AI

Orbital AI’s future is progressing with more API connections for different services and allowing our Agentic AI to be smarter and faster. Although we didn’t get to accomplish everything we wanted with Orbital, we want a bright future ahead for it. The Agentic-AI capabilities are something we will tap into more of, especially for future projects and for Orbital’s success.