DEV Community

Cover image for Vibe Coding 101: I Built A Research Assistant Start-up With v0, Supabase + Grok AI
Dumebi Okolo
Dumebi Okolo

Posted on • Edited on

Vibe Coding 101: I Built A Research Assistant Start-up With v0, Supabase + Grok AI

In the spirit of vibe coding and building something useful without overthinking it, I recently created a lightweight AI-powered web app called ResearchAI Assistant. Doing a lot of my masters dissertation and having to deal with a ton of papers and research, I decided to vibe code something that will be helpful to me if built upon.

It helps researchers, students, and professionals transform complex academic papers into accessible content like slide decks, podcasts, summaries, and visuals—all in just a few clicks.

The front end for this project was already largely built out by a user on v0. I added integrations and UI later on to fit into the project scope.

home page

👉 Check out the GitHub repo


Why I Built This

Academic writing is rich in insight but often dense and hard to share with non-specialists. I wanted a tool that could make research more approachable and engaging, especially for those who need the insights but don’t have time to parse 40+ page papers.

So, I vibe coded this as a weekend project, powered by Supabase for authentication and storage, and a clean UI with responsive pages that feel modern and inviting. It also has the AI aspect running on Grok AI. It’s an idea-first build, not overly polished, but super functional—perfect for demos, hackathons, or real-world use.


Key Features

Here’s what the app currently does:

  • 🔁 Transform Papers into Multiple Formats

transform page

  • 📊 Interactive Slides – Turn technical content into slide presentations with just a paste/upload.
  • 🎙️ Podcast Scripts – Generate AI-narrated audio scripts from research papers.
  • 📄 Executive Summaries – Extract key points into short, digestible formats for stakeholders.
  • 📈 Visual Infographics – Create simplified visual content and diagrams.

    • 📤 Upload or Paste Research
  • Supports PDF, DOC, DOCX, and TXT files (max 10MB).

  • Paste content directly for on-the-fly analysis.

    • 🔐 Secure Auth with Supabase
  • Users must confirm their email before signing in.

email confirmation

  • Email link confirmation workflow is built in for account protection.

How It Works The App Works

  1. Sign up and verify your email
    The app sends you a confirmation email (handled via Supabase Auth).
    verify email address

  2. Upload or paste your paper
    You can drag and drop a document or paste the content directly into the textarea.

upload paper

  1. Choose a content format
    Whether it’s slides, a podcast script, or a summary—you just click a button.

  2. Preview and Download
    After processing, you can preview the output or download it for sharing or presentation.

preview and download


🧠 Behind the Scenes

  • Frontend: Built with React and styled for a dark theme that’s sleek and professional.
  • Backend: Authentication and database powered by Supabase.
  • AI Logic: Content analysis and transformation powered by Grok AI. (I left the actual AI integrations open for future upgrades.)

📸 Screenshots

Here are a few views from the app:

  • 📄 Upload Page Drag-and-drop or paste your document.

upload paper

  • ✅ Analysis Complete Shows estimated reading time, complexity, and key insights.

analysis complete

  • 📊 Examples Page Real-world samples: Machine Learning, Climate Policy, Quantum Computing, and more.

examples page

  • Dynamic Sign-Up page

sign-up page


🔗 Try It Yourself

This is a fun, weekend side project, so feel free to fork it, play with it, or even extend it with other AI models like OpenAI or Claude.

👉 GitHub: github.com/Dumebii/research-assistant


Final Thoughts

This wasn’t about building a “perfect” product—it was about vibing with a useful idea and turning it into a live, testable prototype. If you're a student, researcher, or builder who loves experimenting with AI and interfaces, this is a fun project to explore.
I should also add that I vibe wrote this article. I gathered all the data I need for the article, including screenshots, key points, and relevant links. I fed it into an LLM, and it came up with an article draft, which I went ahead to modify.
The idea is that if I can vibe code an entire app, why can't I vibe write an article draft?

Got feedback or feature ideas? Drop a comment!

Top comments (10)

Collapse
 
juandadev profile image
Juanda Martínez

It looks awesome in the screenshots! But the link to the repo doesn't work :( I want to take a look at it!

Collapse
 
dumebii profile image
Dumebi Okolo

Hiii, thank you so much for catching this.
I don't know why the first link was broken.. But here it is. I have also updated the article.
Thank you.

Collapse
 
harshal_shinde_a16da7de4e profile image
Harshal Shinde

Hi, great work, i would like to take a deeper look into the code but the link to the repo seems to be broken, could you share a working link? Thanks

Thread Thread
 
dumebii profile image
Dumebi Okolo

Hi, thank you so much for pointing this out. I didn't realize the repo was set to private; that's why it kept turning up broken but working on my end. I have resolved that now.

Collapse
 
dotallio profile image
Dotallio

Love this vibe-coded approach! Would be super cool if the podcast feature could generate actual audio files for mobile listening - any plans for that?

Collapse
 
dumebii profile image
Dumebi Okolo

Hmm. I could work on that. I am only being careful not to expend all AI credit.

Collapse
 
werliton profile image
Werliton Silva

Wowww! amazing.

good job.

Collapse
 
dumebii profile image
Dumebi Okolo

Thank you!

Collapse
 
nathan_tarbert profile image
Nathan Tarbert

Pretty cool, I’m always looking for a way to make research easier for myself. Honestly, this is extremely impressive

Collapse
 
dumebii profile image
Dumebi Okolo

Thank you!!