Precursor

Precursor was inspired by a simple idea: “what if our daily digital conversations could make us a little wiser, more creative, and more connected to culture?” We noticed how much time people, especially students, spend on messaging apps, often expressing deep emotions or meaningful thoughts, and wanted to turn those moments into opportunities for reflection and learning. By blending art and technology, we aimed to make culture feel alive and accessible in the flow of everyday life.

We built Precursor using Python, React, Supabase, the OpenAI API, and the Google Custom Search API. The backend detects emotional or thematic cues in user messages using natural language processing, then dynamically retrieves a relevant painting and curates context about the artist, the meaning, and its historical importance. The frontend was designed to be minimal and conversational, ensuring that these insights appear naturally, without interrupting the chat experience.

Throughout development, we learned how to combine emotional intelligence with creative AI outputs to translate human expression into cultural connections. One of our biggest challenges was finding the balance between relevance and subtlety: ensuring the app surfaced meaningful artworks without overwhelming the user. We also had to fine-tune the emotion detection model and optimize API calls for speed and accuracy.

In the end, Precursor taught us that technology can deepen our connection to it.

Built With

• google-custom-search
• openai
• python
• react
• supabase