A Chinese e-reader that surfaced real news articles for language learners.
Earlyworm v1 was the first iOS app I ever built. I was living in Shanghai, studying Chinese, and wanted a better way to bridge the gap between classroom materials and real-world content, so I taught myself Swift and started building. The idea was simple: instead of studying from textbooks, read real Chinese news articles and blog posts with built-in tools to help you through them. It pulled articles from Chinese news sites via RSS, displayed them in a clean reading interface, and had inline dictionary lookup so you could tap any word to see its definition and pinyin.
A flashcard feature helped you review vocabulary from your reading, powered by a Node.js server communicating over WebSockets. A basic hill-climbing algorithm tracked your answers and estimated your proficiency level, adjusting card difficulty in real time.
The app was functional but limited by what was possible at the time: no AI rewriting, no proficiency matching beyond the hill-climbing heuristic, just raw articles with reading aids. Years later, the emergence of generative AI made it possible to revisit the core idea and build what became Earlyworm 2022.
Earlyworm demo showing the article discovery and reading experience.