Builders NYC: Hack Night #1 -- Flex Your Culture!

A parent-child agent to agent specification for managing adhoc interrupt tasks. My main goal was to create a specification. Parent agent/brain owns long term goals and future obligations. It sends interrupts to sub-agents (children) as and when required based on the context going on in the child. The project repo is just a bare minimum MVP which is mostly a UI. I'm gonna create a spec and build on top of A2A protocol.

team name #1

# Culture Remix Translator **Tagline:** Most translation apps translate words. This translates context. ## Problem Standard translation tools convert language, not meaning. Cultural phrases, family habits, traditions, and slang often carry emotional weight, social context, and unspoken care that gets lost in a literal translation. A phrase like "Did you eat?" can sound practical to outsiders while functioning as a love language in many Nepali and South Asian families. People navigating cross-cultural workplaces, friendships, and relationships need a way to explain *why* something matters—not just *what* the words say. ## Solution Culture Remix Translator is a Streamlit app that turns culturally specific concepts into shareable explanations tailored to a target audience and tone. Instead of a single word-for-word translation, it surfaces: - **Literal meaning** — what the phrase or practice says on the surface - **Cultural meaning** — the social, emotional, or relational context behind it - **Common misunderstandings** — where outsiders often get it wrong - **Friend-style explanation** — a warm, personal way to describe it - **Modern analogy** — a relatable comparison for the target audience - **Say it out loud** — a short script someone can actually use in conversation - **Nuance note** — regional variation, sensitivity, and what not to oversimplify The goal is not to flatten culture into something simpler. It is to make culture easier to share without losing the feeling. ## How it works 1. Enter a cultural phrase, tradition, food, slang term, or family habit. 2. Set your background, target audience (e.g. American coworkers), and tone (e.g. first-gen kid explaining to coworkers). 3. Click **Translate the context** to generate a structured explanation. 4. Use the **Shareable explanation** card as a ready-to-say summary. Built-in demo examples include "khana khayau?", Dashain tika, auntie culture, and namaste outside a yoga-studio context. ## Tech stack - **Frontend:** Streamlit with a dark, warm UI - **LLM:** Claude Opus via Portkey / NYU AI Gateway (optional) - **Fallback:** Deterministic demo responses so the app works offline and during live demos without an API key When the live model is unavailable or returns malformed JSON, the app gracefully falls back to curated responses so the hackathon demo always works. ## Why it matters Immigrant and first-generation communities constantly translate context for others—at work, at school, in friendships. This tool gives that invisible labor a shape: clear, respectful, audience-aware explanations that preserve nuance instead of stereotype. ## Demo pitch (60 seconds) 1. Open with: "Most translation tools tell you what words mean. This tells you what they mean emotionally." 2. Click the **khana khayau** example. 3. Set target audience to **American coworkers** and tone to **First-gen kid explaining it to coworkers**. 4. Click **Translate the context**. 5. Walk through literal vs. cultural meaning, the misunderstanding, and the say-it-out-loud version. 6. Close with: "The point is not to make culture simpler. It is to make it easier to share without losing the feeling." ## Run locally ```bash pip install -r requirements.txt streamlit run app.py ``` Optional: copy `.env.example` to `.env` and add `PORTKEY_API_KEY` for live model responses.

shOp

A Chrome extension that recommends culturally-rooted small brands as alternatives to culturally appropriated fast fashion. When you are browsing a product on a big retail site, it surfaces real small businesses selling something similar, with the story behind each one.

Glimpse

Glimpse is a map based app that is a collaborative history of your neighborhood

Team Bão

Bão is a watch party discovery app that helps people find communities through sports. The World Cup is more than a tournament, it's a celebration of culture. Across New York, people gather to cheer for the teams that remind them of home, sharing food, traditions, and unforgettable moments together. But unless you already know someone, it's surprisingly hard to find those communities. That's why we built Bão. Bão means "storm" in Vietnamese, the wave of excitement when an entire room erupts after a goal. We wanted to capture that feeling and make it easy to find. Restaurants, bars, and even households can create a watch party in minutes by sharing their location, the team they're supporting, and what makes their gathering special. AI helps generate a short description, while guests browse an interactive map, filter by team or vibe, and RSVP instantly. We built Bão in one night—two builders, one from Hanoi and one from Karachi. Neither of our national teams made the World Cup, but we know what it's like to look for a place where you belong. The idea starts with the World Cup, but it works for any sport, any city, and any community. Because the best part of sports isn't just the game, it's the people you experience it with.