Why is one English article a day still difficult to sustain?
Learners must find content at the right level, then move between reading, dictionary, translation, saving and review tools. The reading flow breaks, while saved words and sentences rarely return to the learning process.
Target users
- University students preparing for IELTS
- Advanced learners rebuilding reading fluency
- Learners who struggle to find material or sustain a daily habit
Constraints
- Deliver a complete first version without accounts or a cloud database
- Keep reading, lookup and review inside one loop
- Provide layered fallbacks for dictionary and translation services
- Use a 27-article offline sample library while respecting content rights
- Describe product ownership separately from AI code support
Solution
I narrowed the scope to a daily loop: choose a level and length, draw an article, look up and save content during reading, continue through flashcard review and finish with visible check-in feedback.
Key features
Find the right reading
Three IELTS difficulty levels and three article lengths, with an offline sample library for a reliable core experience.
Stay in the reading flow
Definitions, pronunciation, paragraph translation, phrase lookup and sentence saving without leaving the page.
Turn saving into review
Saved words and sentences enter a personal library, with flashcard review organised in sets of ten.
Make consistency visible
A check-in calendar shows streak, monthly and total activity as practical habit feedback.
My contribution
I defined the problem from my IELTS preparation, set the MVP scope, made feature trade-offs, designed the flow, broke down tasks, tested states and invited friends to try the product.
How I worked with AI
Codex and WorkBuddy supported task breakdown, code generation, modification and debugging. I provided context and acceptance criteria, reproduced issues and judged whether each change improved the experience.
Validation and iteration
About ten friends tried the product. Their feedback led to a check-in calendar, flashcard review, punctuation cleaning and layered dictionary fallbacks, plus fixes to popovers, translation availability and streak calculation.
Results
The project has a public product, a complete core flow, early testers, feedback-driven iterations, GitHub history and documented screenshots.
Limitations
It still lacks formal analytics, completion and retention data, accounts and cloud synchronisation. Most data remains in localStorage, and content rights and updating need further work.
What I would build next