f(unc) - a brainrot translator

Inspiration

The internet moves fast—too fast for some generations to keep up. When a Boomer hears "rizz" or "skibidi," they're completely lost. We built f(unc) to bridge the generational gap in internet culture, translating Gen-Z brainrot into language that Boomers, Gen-X, and Millennials can actually understand.

What it does

f(unc) is an AI-powered meme translator that explains internet slang and memes in generation-specific language. Users enter a meme name (like "aura" or "Ohio") and select their generation, and our LangGraph pipeline:

1. Scrapes Know Your Meme for accurate meme data
2. Retrieves generation-specific language context from ChromaDB
3. Uses Claude Sonnet with RAG to generate explanations tailored to each generation's vocabulary and cultural references
4. Displays relevant YouTube videos for visual context

How we built it

AI/ML Pipeline: 3-node LangGraph workflow with RAG (Retrieval-Augmented Generation)

• Scrape Agent: BeautifulSoup4 extracts meme metadata from Know Your Meme
• Curate Agent: Structures meme data and queries ChromaDB for generation-specific language examples
• Explain Agent: Claude Sonnet generates explanations using retrieved context about how each generation communicates

RAG with ChromaDB: We built a vector database of generation-specific language patterns:

• Boomer examples: Formal vocabulary, traditional media references
• Gen-X examples: Skeptical tone, MTV/90s culture, ironic humor
• Millennial examples: Social media slang, workplace references
• Gen-Z examples: TikTok references, internet slang ("no cap," "fr fr," "bussin")

The system retrieves relevant examples for the target generation and injects them into Claude's prompt for accurate style matching.

Tech Stack:

• Backend: FastAPI, LangGraph, Langchain, Anthropic Claude Sonnet, ChromaDB
• Frontend: Next.js 15, React 18, TypeScript, Tailwind CSS, Geist Mono
• APIs: YouTube Data API for video content

Challenges we ran into

• Vector Database Design: Structuring generation-specific context in ChromaDB for efficient retrieval
• RAG Integration: Seamlessly combining retrieved examples with meme data in prompts
• Web Scraping: Know Your Meme's inconsistent HTML structure required robust parsing
• Prompt Engineering: Balancing retrieved context with dynamic meme information
• State Management: Coordinating ChromaDB queries within LangGraph's state flow

Accomplishments that we're proud of

• Built a production-ready LangGraph pipeline with RAG integration
• Designed a ChromaDB schema for storing generation-specific language contexts
• Created prompts that genuinely capture different generational communication styles
• Smooth UI with Claude-style thinking animations and sliding transitions
• Successfully deployed a full-stack AI application with real-time explanations

What we learned

• How to implement RAG with LangGraph for context-aware generation
• ChromaDB's power for storing and retrieving semantic language patterns
• Prompt engineering techniques for style transfer across generations
• Building responsive, animated UIs with Next.js and Tailwind CSS

What's next for f(unc)

• Expand ChromaDB with more generation-specific examples
• Add Gen Alpha and Silent Generation support
• Image-based meme analysis using vision models
• User feedback loop to improve ChromaDB embeddings
• Chrome extension for instant translation while browsing
• Community-contributed language examples

Built With

• beautifulsoup4
• chatgpt
• chroma
• fastapi
• langchain
• langgraph
• next.js
• react
• tailwindcss