🌟 Inspiration
The idea for LocalizeAI was born out of our love for cafés, cozy interiors, work-from-café (WFC) vibes, and, of course, great food! 🍽️ We’re always on the lookout for the perfect place to code, relax, and recharge. But finding that ideal spot—whether it’s a hidden gem or a popular café—often means endless scrolling through TikTok, Instagram, and review sites. 📱
Our vision truly came together after one of our team members attended the MongoDB.local event in Jakarta on August 13, where they introduced vector search capabilities with MongoDB. This sparked the concept of LocalizeAI. Being tech enthusiasts and hackathon regulars, we were inspired to turn this technology into something practical for café and food lovers. With LocalizeAI, we aim to simplify the search for the perfect spots, creating a platform for those who love exploring local hangouts, unique atmospheres, and great food. 🌐🍰
🚀 What It Does
LocalizeAI is a social platform designed to elevate your dining and café experiences in Jakarta. Whether you’re in search of a cozy café, a trendy restaurant, or just a new place to try, LocalizeAI helps you find it with the power of text and image search.
🌟 Key Features:
- 🔍 Effortless Search: Discover your ideal spot with just a keyword or image.
- 💡 Personalized Recommendations: Tailored suggestions that reflect your unique tastes.
- 👥 Community-Driven Reviews: Share experiences, rate spots, and connect with fellow food lovers.
- 📍 Map Contributions: Suggest new places, helping expand our community-driven map of Jakarta’s dining scene.
🛠️ How it Works:
- 🔍 Search: Use keywords or upload an image to find restaurants and cafés that match your criteria.
- 🌐 Discover: Access detailed info, user reviews, and images for each location.
- 📢 Share: Join the community, post reviews, and recommend your favorite spots.
🌈 Positive Impact:
- 👫 Community Building: Cultivates a vibrant community of food enthusiasts sharing their love for local spots.
- 📈 Economic Boost: Supports local businesses by connecting them with more potential customers.
- 🌍 Social Impact: Encourages cultural exploration and a deeper appreciation of Jakarta’s rich food scene.
- 🎉 Streamlined Event Planning: Easily plan gatherings with friends by quickly finding suitable venues.
💻 How We Built It
LocalizeAI was developed with a cutting-edge tech stack optimized for speed and search efficiency:
- Backend: FastAPI for RESTful API services 🚀
- Database: MongoDB Atlas on AWS for vectorized data storage 📊
- LLM Models: Amazon Bedrock Llama3
- LangChain & LangSmith: For managing language model flows and observability 🛠️
- Embedding Model: CLIP-ViT-L-14 for text and image embeddings 🤖

🧗 Challenges We Ran Into:
- 🌐 Data Crawling: Collecting data from diverse online sources was tricky, but we managed to gather detailed info on over 400 coffee shops and restaurants around Jakarta!
- 🖥️ Data Cleaning & Vector Embedding: Processing and embedding the large dataset was resource-intensive, requiring high RAM and GPU power, especially for image embeddings.
- 🔄 Model Integration Reliability: Keeping integrations smooth was challenging, as Amazon Bedrock sometimes encountered errors or became temporarily unavailable.
🏆 Accomplishments We’re Proud Of
- 📍 Extensive Data Collection: Compiled a robust dataset of over 400 coffee shops and restaurants across Jakarta, with detailed information to enhance discovery and user engagement.
- 🔍 Implementing Vector-Based Image Search: Successfully integrated vector-based image search using MongoDB Atlas, allowing users to discover places by uploading photos, enhancing search precision and user experience.
- 🖥️ Efficient Vector Embedding: Achieved high-quality text and image embeddings to power accurate, relevant search results, optimized to handle large datasets smoothly.
- ⚡ Smooth API Integrations: Seamlessly integrated Amazon Bedrock, and MongoDB Atlas, creating a powerful backend to support fast and reliable AI-driven recommendations.
- 🌐 Progressive Web App (PWA): Built LocalizeAI as a PWA, providing users with an app-like experience directly from their browser, eliminating the need for downloads and making access seamless on any device.
- 🌱 Community-Driven Platform: Created a community-focused design that encourages users to share reviews, recommend spots, and contribute to the growing database, fostering a vibrant, authentic food-lovers community.
🎓 What We Learned
- The importance of user feedback to refine features and improve user experience 💬
- Insights into optimizing MongoDB for high-performance vector search 🚀
- A deeper appreciation for community-driven development, realizing the value of user-generated insights 🙌
🔮 What’s Next for LocalizeAI
- 🌍 Expanding Data Coverage: We’re set to grow beyond Jakarta, starting with popular Indonesian cities like Bandung, Surabaya, and Bali. Our longer-term goal is to broaden our reach across Southeast Asia, making LocalizeAI a go-to resource for culinary discovery throughout the region.
- 🤝 Partner with Local Businesses: We aim to collaborate with cafés, restaurants, and local food brands to offer exclusive promotions, events, and deals to our users. These partnerships will also support local businesses by increasing their visibility within our platform.
- 🎶 Collaborative Playlists: Users will soon be able to save, curate, and share lists of their favorite places, perfect for group recommendations or planning outings with friends.
- 💡 Enhanced AI Features: We’ll be adding more intelligent, personalized recommendations that adapt to each user’s preferences and past interactions, making discovery even more engaging and relevant.
- 📈 Deeper Community Engagement: Interactive maps, customized lists, and community-driven features will let users connect with each other and find even more of Jakarta’s hidden gems.
🌟 Vision for LocalizeAI
Our vision for LocalizeAI is to be the leading culinary exploration platform across Southeast Asia—a companion that empowers users to easily discover unique, local dining experiences. By building a connected community of food lovers, supporting small businesses, and sharing authentic food experiences, we aim to transform how people explore and appreciate local flavors, one city at a time. 🍲🌏
Built With
- amazon-ec2
- bedrock
- clip
- docker
- embedding
- fastapi
- langchain
- llama
- mern
- mongodb
- mongoose
- nestjs
- node.js
- pwa
- python
- react
- redis
- s3
- swagger
- typescript
- vector-search





Log in or sign up for Devpost to join the conversation.