StudyFlow Agent

StudyFlow Agent is a local-first education product for students who are trying to stay on top of classes without turning their week into another admin project. It takes messy goals, assignment deadlines, class notes, prep work, and blockers, then reshapes that chaos into a weekly plan that feels structured enough to follow and lightweight enough to keep using.

Inspiration

Students already have the ingredients for planning. They have lecture notes, assignment briefs, exam dates, club responsibilities, and a running stream of things they know they should do next. The problem is that all of this information lands in scattered tabs, messages, notebooks, and half-finished checklists. StudyFlow Agent was built around that reality. Instead of asking students to rebuild their week inside a heavyweight tool, it starts from the mess they already have and turns it into a usable plan.

What it does

StudyFlow Agent converts freeform academic input into prioritized tasks, a Monday-to-Friday study roadmap, a live execution board, and an exportable action brief. The plan stays adaptive because the system re-scores the work around urgency, impact, and effort whenever the student regenerates the workspace. The result is not just a prettier to-do list. It is a weekly operating layer for schoolwork, deadlines, and accountability check-ins.

The exported brief is especially useful because it gives the student something reusable. They can copy it into a study session, send it to an accountability partner, or hand it to another agent as a clear next-action summary.

How we built it

The product stays intentionally practical. StudyFlow Agent is a static web app built with HTML, CSS, and JavaScript. Planning logic, state management, export generation, and the demo workspace live in the browser. Local persistence is handled with localStorage, and a service worker keeps the shell available offline. That means there is no backend, no database, no auth layer, and no secrets required to run the product.

The project also includes automated tests for planning and brief generation, a repeatable build process that outputs to docs/, and a clean asset pipeline for screenshots, narration, and the final walkthrough.

Challenges we ran into

One challenge was making the experience feel agentic without hiding behind a backend or vague AI claims. The product needed to stay local-first and static, but it also needed to produce outputs that felt smart, structured, and immediately actionable. That pushed the planning layer to do more with ranking, grouping, and summarization inside the client.

Another challenge was presentation. Education products often become visually bland or overly utilitarian. We wanted the experience to feel polished enough for a hackathon demo while still looking credible for real student use.

Accomplishments that we're proud of

• Turning messy student input into a full weekly workflow instead of a single generated list
• Keeping the architecture fully static and local-first without losing product clarity
• Building an export flow that makes the output useful beyond the app itself
• Creating a polished demo surface that is credible for judges and practical for actual students

What we learned

We learned that an agentic experience does not need a complicated stack to feel useful. If the product starts from the user’s real mess and returns a plan that is structured, accountable, and easy to reuse, it already creates meaningful leverage. We also learned that local-first constraints can improve focus because they force the product to be faster, simpler, and easier to verify.

What's next

The next step is to make the planner more personalized around course load and study style. That includes recurring class templates, better blocker tracking, richer accountability prompts, and cleaner handoffs for tutoring, peer study groups, and future student-support agents. The core direction stays the same: keep the product lightweight, local-first, and genuinely useful during a real school week.

Built With

• css
• html
• javascript
• localstorage
• service-worker