Cultural Intelligence Assistant

• 

  Frontend

• 

  Frontend

• 

  Frontend

🚀 About the Project

💡 Inspiration

We were inspired by a key challenge many creators face: how to tailor content ideas that truly resonate with a specific audience. In a world full of generic AI tools, we wanted to go deeper—bridging real cultural insights with generative creativity.

🛠️ How We Built It

Our system combines:

• 🧠 Large Language Models (LLMs) to understand the creator’s intent and parse prompts into structured insights.
  
• 🌐 Qloo’s cultural graph API to identify real-world audience preferences (e.g., music, games, places).
  
• 🧩 A custom pipeline to validate, clean, and enrich responses before returning tailored creative recommendations.
  
• 🌍 React (for demo) and a JavaScript frontend (for production) that call our backend hosted on Hugging Face Spaces.

We allow users to input any natural-language description of their target audience, and we return curated, culturally relevant inspiration to build better content.

⚙️ Technologies Used

• Python (FastAPI, Langchain)
  
• JavaScript (React)
  
• Hugging Face Spaces (API hosting)
  
• Qloo API (cultural insights)
  
• OpenRouter / OpenAI (LLM)
  
• Docker
  

🔍 What We Learned

• Parsing natural language into structured signals that match Qloo’s API needs required careful prompt design and robust validation.
  
• Combining symbolic APIs like Qloo with probabilistic AI like LLMs is incredibly powerful but tricky to coordinate.
  
• LLMs are great at adapting user input—but require guardrails to avoid malformed outputs when integrating with strict APIs.

🧗 Challenges

• Designing prompts that consistently produce clean, structured output for downstream parsing.
  
• Handling edge cases where users don't specify enough information (age, interests, location).
  
• Deploying a reliable and responsive API that integrates multiple services (LLM, Qloo, frontend) under time pressure.

💬 Final Thoughts

We believe this tool can become a powerful assistant for marketers, storytellers, and creators—anyone who wants to understand their audience better and build content that matters.

Built With

• fastapi
• huggingface
• javascript
• langchain
• netlify
• python
• react