The Metaphor
Eddy is named after the circular currents that form in water — patterns that emerge naturally from flow. Product feedback works the same way. Individual comments seem random, but when you look at the currents, patterns emerge. Feature requests cluster. Pain points repeat. Opportunities surface.
Eddy captures feedback from every channel your team uses and applies AI classification to reveal those patterns.
AI With Privacy
Here’s the tension most feedback tools ignore: AI-powered classification is genuinely useful, but it typically requires sending your customer data to third-party AI services. Those services have their own data retention policies, training practices, and privacy standards — none of which you control.
Eddy is designed to run AI classification on infrastructure you control, using models that don’t retain your data for training. The AI serves you, not the other way around.
The Architecture
React 19 and Vite power a fast, modern frontend. Hono provides a lightweight API layer. PostgreSQL stores feedback data with encryption at rest. Redis handles real-time features and caching. The stack is deliberately simple — fewer dependencies mean fewer places for data to leak.
For Product Teams
Eddy is built for product teams who take customer feedback seriously but don’t want to compromise on privacy. Every piece of feedback your customers share is a trust signal. Eddy helps you honor that trust while still extracting the insights you need to build better products.