facial recognition
past conversation lookup

Inspiration

Secondbrain started from a simple frustration: the most meaningful conversations in our lives still happen offline, but all of our tools assume everything lives in apps. We’d meet people on campus, make promises (“I’ll send you the notes”, “I’ll intro you to X”, “Let’s meet next week”), and then rely on memory and scattered notes to follow through. By the time the next interaction came around, the context of who someone was, where we met, and what we owed them had already faded.

We wanted an “AI second brain” that actually understands our real lives, not just our docs and chats. Secondbrain was inspired by the idea of structured relationship memory: remembering real interactions with real people, surfacing open loops before the next conversation, and helping you show up prepared instead of scrambling to reconstruct the past.

What it does

Second brain runs during conversation, captures peoples faces, promises, transcripts, creates profiles for friends, looks them up using facial recognition. We turn that conversation into structured memory tied to a specific person, time, and place. For each person, it builds a running profile of topics, promises, and context across multiple interactions.

Before you meet someone again, Secondbrain can surface what matters: where you met last time, key topics, and any open loops you still owe them (for example: “send CS160 notes”, “share YC contact”, “hit them up when they’re in San Diego next month”). Instead of digging through messages or trying to remember everything, you get an at-a-glance relationship overview so you can follow through consistently.

How we built it

We built Secondbrain as a web app powered by facial recongition, underlying memory and a contexct search layer(NIA).

Challenges

We wanted to make this a wearable but ran into challenges with Meta Glasses (We're NGMI)

Longterm

Secondbrain is meant to be a wearable, we want to expand to meta glasses when they open up a good livestreaming SDK so you can discreetly save all your interactions (all data is stored locally, we do not process anything)!