mymento

Inspiration

We were inspired by two distinct gaps in how families connect:

The "Awkward Silence" Problem:

• Intergenerational conversations often stall because young people don’t know what to ask, and older generations struggle to articulate deep memories without gentle guidance.

The "Lost Context" Problem:

• Even when stories are told, the younger generation lacks the historical scaffolding to truly understand the significance (e.g., why a specific year was "hard").

We asked: What if an app could act as both a skilled conversation host in real-time AND a historical editor afterward? MyMento is the first platform that facilitates the deep talk as it happens (via AI prompts and flow redirection) and then transforms that talk into an immersive "Wrapped" reflection (via archival media and context). It’s not just a recorder; it’s a active bridge across generations.

What it does

MyMento operates in two equally vital modes:

Mode 1: In-Situ Conversation Facilitation (The "Host")

Smart Prompt Engine: Generates culturally-aware, open-ended questions in real-time based on the current topic (e.g., “You mentioned moving to Poland; what was the first thing you bought with your own money there?”). Flow Redirection: Detects when the conversation stalls, loops, or becomes too superficial using sentiment and topic modeling, then gently nudges participants with: “Oliver, ask Van about the smell of that first apartment,” or “Van, can you tell us about a time you got stuck in a storm?” Language Scaffolding: Live translation hints and simplified phrasing suggestions for the younger user if the older user’s dialect or broken English becomes a barrier, ensuring the flow never breaks. Deep-Dive Triggers: Identifies vague mentions (“That bad year”) and instantly prompts for specifics: “Which year was the 'bad year'? Was it the drought or the winter?”

Mode 2: Post-Conversation Reflection (The "Wrapped")

Archival Immersion: Automatically pulls real photos, news reels, maps, and weather data tied to the exact time/place mentioned during the session. AI Context Cards: Generates neutral, 1–3 sentence historical explanations for key stories (e.g., “1960 Fujian drought: Rainfall 30% below average, leading to widespread crop shortages”). Cinematic Recap: Produces a 60–90 second “Wrapped” video syncing the speaker’s voice with archival visuals and emotional theme clusters. Shareable Heritage: Exports highlights as social reels or private family microsites, turning the session into a permanent, discoverable archive.

How we built it

Claude Code -> TypeScript Next-JS app.

Built With

• claude
• next-js
• typescript