Echo AI: A Personal Voice Journal

Inspiration

"Adjusted for the subjective increase in how fast time passes, life is half over by 23 or 24. Don't waste time"

Something everyone in their 20's experiences is the "slipping of time". We blink our eyes and a year has passed by. We have no idea how the days, months and years pass. The ONLY way you bring structure into this chaos is with data - your life data. But typing/writing every single day into a journal is ridiculously boring - no one to prompt me what to talk about, no dynamism - it's just super hard to be consistent and track that "life data".

What it does

Echo AI is a personal voice assistant that you can talk to every night before you sleep. Once you finish talking to it, it:

Creates a "mood score" for the day
Adds important details covered in the session to its 'long term memory'
Adds important events in your life
AND THEN REMEMBERS ALL THIS DATA THE NEXT TIME YOU TALK TO IT - So it remembers how you've been feeling this past week, what's been happening in your life and what your triggers are.
It also uses scientifically proven methods to deal with situations where user is anxious, depressed, or overwhelmed - The voice agent dynamically decides what exercise to show the user based on the situation - The list of exercises present are: Breathing exercises, behavioral activation quest, grounding exercise and progressive muscle relaxation

How we built it

We first ideated and decided on an idea that everyone liked (was surprisingly hard) - 2 of us created the PRD, while the other 2 built the base of the project and figured out nuances about the eleven labs conversational agent. Once we had a plan and knew the roles each one had to take up - we did what we did best - Build. And we built through the night.

Challenges we ran into

The main challenge we ran into was the fact that once eleven labs called a client tool, there was no option where we could tell the agent to "pause" while the user interacted with that tool. These "tools" were something that the user interacted with in the UI, so there was no way to know how long the user would take before he comes back to the agent.

The agent continued to keep talking even though we tried prompting it to keep quiet after a tool is called. We found a hack by increasing the 'turn timeout' of the agent, which works as a temporary solution, but isn't perfect because the agent would then have no way of knowing whether or not user is still playing the game.

Accomplishments that we're proud of

Within less than a day, we:

Decided on an idea
Created an entire PRD (though all of us come from dev backgrounds)
Did thorough research on the scientific methods used
Finished 90% of the features we'd hoped to finish
Built a FANTASTIC UI

What we learned

Our greatest takeaway from this hackathon is that it's not super easy to work with teams. You have to make sure that you take into consideration every single member's opinion. Taking into consideration that we just had 4 members, we finally understood why so many startup leaders tell that "co-founder quarrels" are why most startups fail.

What's next for Echo AI

We plan on putting the app up behind a paywall and going live within the next week. We KNOW this app is going to help people, and using the twitter reach that all 4 of us have, we plan to market the product and get it into users' hands ASAP. Once that's done, it's going to be:

Take customer feedback -> iterate till we find product market fit. We're very excited to launch Echo!