FinaMeter

Master financial literacy through AI-powered quizzes on real market news.

Comment

Inspiration

Fast-moving domains generate more content than any individual can process, and most feeds optimize for volume rather than learning. We wanted a system that converts “keeping up” into a repeatable habit where users actively check understanding instead of passively scrolling. ​

What it does

FinaMeter turns real-time market news into structured learning: it ingests fresh articles, produces AI summaries/tutorial-style explanations, and attaches short multiple-choice quizzes with feedback. Users earn XP, maintain streaks, and track mastery with sector-specific gauges that reflect recent engagement (not just historical activity). ​

How we built it

We built FinaMeter as an automated pipeline: articles are pulled from multiple sources, cleaned into usable text, and then processed by an LLM step that generates consistent, app-ready learning artifacts (summaries, sector tags, and quiz questions). A scheduled backend orchestrates ingestion, deduplication, and batch processing so new content appears continuously without manual curation. We then wrapped this pipeline in a product loop—read → quiz → feedback → progression plus real-time market context and social competition to keep learning engaging over time. ​

Challenges we ran into

The hardest part was getting reliable, structured quiz output from an LLM across diverse article styles while keeping latency and failures under control. We also had to make leaderboard and progression features fast without recalculating heavy aggregates on every request, and to balance frequent updates with external API limits during market hours. ​

Accomplishments that we're proud of

We’re also proud of the production-minded learning loop we built around application, not just consumption: users can test understanding with per-article quizzes, then apply it in our daily market prediction game (UP/DOWN picks on a rotating stock pool with automated resolution and XP rewards). The Sector Gauge system adds a second, more honest progress signal-each favorited sector has a 0-100 score that rises with accurate quiz performance but decays when new content in that sector goes unread, so mastery reflects recent engagement rather than only accumulated activity. ​

What we learned

We learned how to design a full-stack product around automation: scheduling, deduplication, batching, and failure handling matter as much as UI. We also saw how much product clarity improves when progress metrics reward consistency and understanding rather than raw consumption. ​

What's next for FinaMeter

We plan to expand beyond finance by plugging the same ingestion + generation pipeline into other fast-moving domains (e.g., AI policy, crypto regulation, geopolitics). Next steps also include deeper personalization (better feeds and revision quizzes), stronger analytics in weekly reports, and scaling the background processing into dedicated workers as usage grows.

Built With

Share this project:

Updates