ZapVite – Conversational EV Charger Locator

Inspiration

Finding an EV charger shouldn’t feel like hunting for buried treasure. Most apps show static maps, but drivers need something faster and more intuitive — especially on the go. We wanted to create a conversational assistant that feels like asking a friend:

“Where can I charge up right now?”

Instead of digging through menus or filters, EV drivers deserve a quick, natural way to find charging stations.

What it does

Zapvite is an AI-powered chatbot that helps EV drivers locate chargers in real time using natural language. Users can simply type queries like:

• “Where’s the nearest fast charger?”
  
• “Find a CCS charger near my office.”
  
• “Are there any free chargers at the downtown library?”
  

The chatbot intelligently filters by:

• Connector Type: NACS (Tesla), CCS, CHAdeMO
  
• Charging Speed: Level 2 (AC) or Level 3 (DC fast charging)
  
• Network: Electrify America, EVgo, ChargePoint, etc.
  
• Cost: Paid vs. free chargers
  

How we built it

• Frontend:
  
  • Built with React and react-chatbot-kit for a clean conversational UI
    
  • Auto-growing chat input with Shift+Enter for multi-line queries
    
• Backend:
  
  • FastAPI (Python) handles incoming queries
    
  • CORS-enabled API for secure cross-origin requests
    
• Data & AI:
  
  • CSV dataset of charging stations
    
  • Lightweight semantic search pipeline to match user intent
    
  • Query memory to allow follow-up conversations (“What about free chargers instead?”)
    
• Extra capabilities:
  
  • Location inference using browser geolocation APIs
    
  • Modular backend to integrate real-time charging APIs in the future
    

Challenges we ran into

• Natural language complexity: Users phrase queries in many different ways, making parsing tricky.
  
• Session memory: Getting the bot to remember previous context required careful state management.
  
• Location access: Prompting for geolocation while keeping UX smooth and privacy-friendly.
  
• Hackathon time constraints: Building and integrating everything end-to-end quickly.
  

Accomplishments we’re proud of

• Built a fully functional chatbot that actually finds chargers instead of showing a static map.
  
• Created a modular backend that can be extended with real-time data feeds.
  
• Delivered a polished UI with thoughtful UX touches that make the bot feel natural to use.
  

What we learned

• Combining LLMs with structured datasets gives focused, trustworthy answers.
  
• Small UX improvements (auto-growing text boxes, better message wrapping) make a huge difference.
  
• Balancing AI flexibility with reliable filter logic is critical for user trust.
  

What’s next for Zapvite

• Live charger availability: Real-time integration to show if a charger is free or in use.
  
• Integrate Map and Direction navigation: Allows users to navigate to the EV charger location directly through our app
• Voice input: Hands-free interaction for drivers.
  
• Personalization: Using session memory to tailor results over time.
  
• Mobile-first deployment: As a PWA or native app, and possibly in-car integration.

Built With

• fastapi
• javascript
• kaggle
• python
• react