DM if Interested

Top 5 matches from real builders, already building. Scraped, filtered & scored for skills, intent & founder energy. Skip the chase. Build with people who get it.

Comment

DM if Interested — Matching Cofounders with Speed & Ease

TL;DR
Built out of personal frustration with traditional cofounder platforms, DM if Interested helps early-stage founders match with high-signal cofounders based on public intent, skills, and vibe, using smart scraping, automation, and LLM-powered analysis. Built solo with zero coding background, using tools like Apify, Supabase, n8n, and Deepseek.

🌱 Inspiration

When I started this hackathon, I was planning to build something totally different, a long-time passion project. But it proved too complex. Around the same time, I was working on my thesis: a platform to improve collaboration in multicultural student teams. I’d just learned I needed a cofounder to apply for an incubator to fund and bring that thesis to life, but finding one turned out to be nearly impossible.

Most cofounder matching platforms felt… off. Ghost profiles. Clunky forms (which i sadly didn't have time to fix through my product, but soon...). Manual searches. Some even felt like they were built only for Stanford grads and Y Combinator alumni. None of them felt made for people like me a scrappy, design-driven, and just getting started.

That’s when it clicked, my thesis project already included a smart team-matching system. What if I adapted that logic to help people find cofounders they actually vibe with?

🤝 What It Does

DM if Interested matches solo builders with potential cofounders by scraping real public posts from platforms like Reddit and using LLMs to detect intent, skills, and “founder energy.”

It generates curated matches based on live user needs, not just static filters, and gives builders a jumping-off point for high-signal outreach, no cold DMs or endless browsing required.

🛠️ How I Built It

I started building this new project in mid-June — with zero coding experience and no confidence in my "vibecoding" abilities either. I forced myself not to start with design (which is usually my default), and instead jumped straight into Bolt to build the skeleton.

Within an hour, I had a working landing page, and for the first time in weeks, I felt momentum.

Then came the hard parts.

I started building this in mid-June with zero coding experience and only basic familiarity with no-code tools. I skipped design (for once!) and jumped straight into Bolt to build the skeleton.

From there, I stitched together a full MVP using:

  • Supabase for the database
  • 🕸️ Apify to scrape public Reddit posts (via a student-supported plan)
  • ⚙️ n8n to build and automate the backend workflows
  • 🧠 ChatGPT + Perplexity to assist with code, logic, and prompt engineering
  • 🧠 Deepseek as the LLM to analyze and filter scraped post content

The current version scrapes Reddit for relevant cofounder posts, feeds them through prompts, and returns potential matches. And… it works! I’ve already found real leads I plan to reach out to.

🧗‍♀️ Challenges I Ran Into

  • Steep learning curve: I had to learn tools like n8n from scratch, it took a full day just to get one node working
  • Back end problems: Built end-to-end with no-code tools like Supabase, Apify, and n8n — which meant a lot of time debugging automation workflows. Some flows are still finicky, so the live version includes a demo profile in case matching doesn't run perfectly in real time.
  • Scraping concerns: Managing APIs, exposed keys, and legal concerns
  • Bolt UI quirks: Once I built the app structure, Bolt resisted design edits, and took a long time to make changes which forced me to settle for the user experience initially created. But some "vibecoded" happy accidents made it through
  • Data limitations: Reddit is my only data source so far, adding more platforms required more time and funds
  • Prompt design: Creating meaningful matches via prompt logic needs lots of iteration, which I didn’t have time and budget for

🏆 Accomplishments That I'm Proud Of

  • Built a functioning MVP solo with no coding background
  • Learned and deployed n8n, Supabase, and Apify
  • Created a working automation that outputs real leads from live data
  • Shifted from a failed idea to a functioning product mid-hackathon
  • Validated a problem I personally face, and started solving it

📚 What I Learned

  • Function > polish (especially early on)
  • Backend logic and scraping are just as creative as visual design
  • AI-powered matching has huge potential, if prompts and automations are tuned well
  • Even low-volume MVPs can be surprisingly useful if they're targeted and intentional
  • Building in public platforms (Reddit, Twitter, etc.) offers way more insight than just relying on user input forms

🔧 Key Features

  • Conversational Onboarding (Supabase)
    Uses Supabase for auth, profile data, and user onboarding. Each user’s intent, skills, and idea are stored in Supabase’s Postgres DB, forming the match criteria.

  • Instant Matching via AI + Workflows
    Onboarding data is processed through a custom n8n automation (self-hosted on Render) that scrapes relevant posts from Reddit and compares user needs to post context using embedding-based filtering.

  • Dynamic Match Display (Supabase + Netlify)
    Final match results — including reason for match, compatibility matrix, and message prompts, are fetched from Supabase and displayed on a smooth Next.js interface hosted on Netlify.

  • Idea Board for Technical Builders
    A curated list of startup ideas uploaded by founders, filtered by quality signals, allows technical users to find meaningful projects to join, solving the “where do I find good ideas?” pain point.

  • Custom Domain Integration
    The project is live at a custom domain, registered via Entri, aligning with the bonus prize for branded domain usage.

🎯 Challenge Relevance

  • 🔵 Supabase Challenge
    Supabase powers the full backend of the app: authentication, user data storage, matching logic, and the idea board.

  • 🟣 Netlify Challenge
    The frontend is deployed on Netlify.

  • 🟡 Custom Domain Challenge
    The app uses a clean, memorable custom domain registered via Entri, enhancing branding and user trust.

🚀 What’s Next for DM if Interested

  • Add more public data sources: Product Hunt, Twitter, Polywork, and more
  • Improve prompt logic and filters to surface more accurate matches
  • Experiment with user onboarding flows to personalize results and make onboarding feel easy
  • Strengthen privacy + security, including key management and safe scraping (might need a cofounder here :))
  • Explore monetization options
  • Test matches with real users and gather feedback to improve relevance and UX
  • Refactor workflows for speed, stability, and modularity

⚠️ Heads up for the judges:

Due to fragile webhook automation (n8n), the live app might not respond instantly or at all right now. But it does work, as shown in the demo.

👀 For a working preview of the match experience: • Use the demo login below • Or email me for a live walkthrough: ruthasfawint@gmail.com

Demo login: Email: zedecreates@gmail.com Password: 12345678

Thank you for understanding! This was built fast and solo, the logic does work — even if the infrastructure is still a bit shaky.

This project started as a personal pain point. Now I believe there’s massive potential in a cofounder matching platform built not just around skills, but around drive, goal, and even personality.

Built With

  • apify
  • bolt.new
  • chatgpt
  • deepseek
  • n8n
  • preplexity
  • supabase
Share this project:

Updates