Catalog Copilot is a web-based product discovery experience built with Flutter Web and Algolia as part of the Algolia Agent Studio Challenge.
The project focuses on delivering a consumer-facing, non-conversational copilot that improves product search through relevance tuning, structured data, and low-latency retrieval — without relying on chat-based interactions.
🔗 Live Demo:
https://algolia-agent-challenge.web.app
- 🔍 Fast product search powered by Algolia
- 🧠 Copilot-style insights without conversational UI
- ⚡ Low-latency search experience
- 🏷 Quality Score (QS) visualization
- 📊 Product detail view with contextual signals
- 🎯 Search-driven UX focused on decision support
- 🌐 Built with Flutter Web
Catalog Copilot acts as a silent assistant that enhances product discovery by:
- prioritizing relevant results
- surfacing product quality signals
- reducing search friction
- guiding user decisions in context
Unlike chat-based assistants, this approach keeps users in flow while relevance does the heavy lifting.
- Flutter Web
- Dart
- Algolia Search
- Indexing
- Searchable attributes
- Relevance tuning
- Dio (HTTP client)
- GoRouter (navigation)
- dart-define (secure environment config)
- Clean separation between UI and search layer
- Search service abstraction for Algolia
- Typed product model with Freezed
- Responsive layout with reusable UI components
- Production-ready environment configuration
This project uses flutter_dotenv.
cp .env.example .env
flutter pub get
### 1️⃣ Clone the repository
```bash
git clone <your-repo-url>
cd flutter_challenge