ResearchFlow AI: Helping Students Navigate the Research Maze
This is a submission for the World's Largest Hackathon Writing Challenge: Beyond the Code.
Why We Built ResearchFlow AI
Starting a research project is overwhelming.
As students, we've all faced that paralyzing moment when you know you need to "pick a topic," but every path feels too broad, too narrow, or already explored.
During the hackathon, our team decided to tackle exactly that: the initial, chaotic phase of academic research. We built ResearchFlow AI, an AI-powered assistant designed to help students move from "I’m interested in renewable energy" to "Here’s a focused research question, key papers, and a rough outline to get started."
We knew we wouldn’t have time to build an entire academic writing platform, so we intentionally reduced the scope to focus on the first steps—the ones that cause the most stress and often stop students before they start.
The Problem We Addressed
Most students face three huge problems at the start of a research project:
- Choosing a Topic: It’s hard to narrow down a broad interest into something specific, original, and researchable.
- Finding a Gap: Even if you pick a topic, identifying a genuine research gap is intimidating and time-consuming.
- Handling Information Overload: There’s too much literature out there, and most tools don’t help students find or summarize what actually matters.
What We Built: The ResearchFlow AI MVP
✨ Key Features
AI Research Gap Identifier & Topic Refiner
Input a broad idea, and the AI suggests narrower sub-topics and potential research gaps based on recent academic trends.Curated Literature Collector & Summarizer
The AI fetches a small, highly relevant set of academic papers (5–10) and provides concise, 2–3 sentence summaries of each one. It highlights how each paper relates to your research gap.Basic Outline Builder
Drag and drop papers or insights into a simple outline (e.g., Introduction, Lit Review Sections). The AI suggests section headings based on the collected literature.
What We Didn’t Build (On Purpose)
To keep the hackathon scope realistic, we focused only on the core research pain points.
We skipped features like:
- Grammar/style assistance
- Full reference management
- Supervisor communication tools
- Stress or time management features
- Collaboration or co-writing tools
We wanted to keep our prototype tight and impactful.
Tech Stack Snapshot
- Frontend: React, TailwindCSS, Shadcn-ui
- Backend: Astro
- APIs: Perplexity
- Database: Supabase's PostgreSQL
How We Measured Success
During the hackathon, we kept asking:
- Can a user go from a broad idea to a specific research question with AI assistance?
- Do the literature suggestions make sense and save time?
- Is the outline builder intuitive and useful?
Early tests said yes to all three, and that’s what kept us motivated.
Reflections: Beyond the Code
Hackathons aren’t just about shipping fast—they’re about collaboration, creative constraint, and community learning.
Working on ResearchFlow AI reminded us of the power of focusing on student pain points instead of just features. We hope this project sparks more conversations about using AI to make academic research less stressful and more accessible.
Shoutouts & Thanks
Big thanks to the mentors, organizers, and fellow hackers who shared feedback and ideas during the event!
If you're reading this and you're a student who’s ever felt stuck at the “Where do I even begin?” phase of research—we see you.
Hopefully, tools like ResearchFlow AI can help make that first step easier.
Top comments (0)