In this episode of the Azizi Podcast, host Samir Azizi is joined by developer Kirill Spitsyn to explore the development journey behind Bespoke Beaver — a fun and functional AI-powered app created by the Toronto DAO team. This tool scrapes your Twitter persona and recommends 5 tailored locations in Toronto based on your online traits, using Exa and ChatGPT.

🔍 In this episode:

- Exa vs ChatGPT – performance, reliability, and use cases
- Inside the 3-Hour Dev Challenge that kickstarted the app
- Tech breakdown: Exa web scraping, ChatGPT for persona modeling, Google Maps API for location plotting
- Challenges with AI accuracy, costs, and output formatting
- Backend strategy: caching, defaults, and balancing speed with UX
- Real-world use cases for conferences like Futurist & Consensus Toronto
- Future plans: city expansion, gamification, and monetization through local partnerships

🛠️ Stack Used:
- Exa – AI web scraper for Twitter content
- ChatGPT (OpenAI) – generates user persona + location recommendations
- Google Maps API – interactive map visualization
- Simple backend glue logic to orchestrate flow & fallback

💡 Whether you're an AI enthusiast, indie hacker, or dev building scrappy MVPs, this episode dives deep into the practical and creative side of AI app development, persona modeling, and working lean with limited resources.

📍Try it out at upcoming Toronto events or just for fun: Bespoke Beaver gives AI-personalized location tips based on your X presence: https://bespokebeaver.tdao.to/
Follow Toronto DAO on X: https://x.com/dao_toronto Connect with Kirill on LinkedIn: https://www.linkedin.com/in/kirillspitsyn/ Connect with Samir Azizi: LinkedIn: https://www.linkedin.com/in/samirazizi X: https://x.com/adoublezi 🎧 Subscribe on Spotify: https://open.spotify.com/show/3FDd0dQi8vvTnQz5h5QwAz