Redcart

• 

  A sample demo of an account search

• 

  The landing page

• 

  The form for creating a new report in the reported account's page

• 

  The demo of a reported account's page

• 

  More on the reports page

Inspiration

I heard from about 3 of my friends about the one time they had ordered something from a social commerce store and got scammed. Then I started to intentionally ask most women because it was more common in women. Surprisingly, almost all of them had a similar experience. Then, just like the red car experiment, the scam stories started popping up all over social media. On researching, I found that social commerce scams are surging. In 2025, nearly 30% of scam losses originated on social media, totaling $2.1 billion On talking to and observing some victims, I also learned that:

1. The victims weren't willing to stop shopping from social e-commerce(mostly Instagram and TikTok). The argument was how unique some deals and items were. But they always swore that they were going to be more careful next time.

2. The victims scammed always shared the story freely, even on social media, with the intention of warning others, but that is not effective if one does not have a big following to make it effective, and if they do, the story will be faced out by the algorithm, and it can only reach enough people before then. So in the end, the stories of coution went to waste.

3. As much as most victims swore to be more careful, it was hard since there is no structure for checklists, and the scammers keep changing tactics.

4. It is a global problem that is not solvable by reliance on the social media platforms themselves.

5. A tool for a quick lookup would be helpful, and the bitterness of being scammed would keep it updated, in the effort of trying to get back at the scammer, which eases the helplessness too.

What it does

RedCart is a community-driven scam reporting and verification platform where shoppers can report fraudulent sellers on TikTok, Instagram, Facebook, and private e-commerce stores. Before buying from an unknown seller, users search for their username to see if others have flagged them. Each report adds a "redflag" to the account, creating a collective early-warning system for online shoppers.

How we built it

I started by describing the core problem and the MVP feature set: search, browse, account detail with red flags, report creation with image upload, community voting, and commenting. MeDo scaffolded the entire React + Vite + TypeScript + Tailwind CSS project with shadcn/ui components and Supabase backend integration. I then iterated with MeDo to refine the dark theme (#ff4500 primary, #00baff accent), add the demo data seeding, implement username-only authentication (using synthetic emails since Supabase Auth requires email), and build the admin dashboard with real-time stats. Each conversation was built on the previous one. Starting from the project scaffold, to page layouts, to auth logic, to polishing and bug fixes.

Challenges we ran into

Medo creates the apps in React by default and allows them to be previewed on the platform. I tried to port the app to Django and make it preview, but it did not work despite trying a few workarounds. So I had to roll back to React at some point and focus on getting the demo correct first.

Accomplishments that we're proud of

The demo account system has realistic fallback reports. When a seeded account has no real reports yet, the detail page automatically generates three detailed, believable demo reports with varied scam categories (Non-Delivery, Fake Product, Payment Fraud), specific dollar amounts, and simulated comment threads from fake users. Combined with the 12 pre-seeded demo accounts across all platforms, the entire user journey from search to detail page is instantly demonstrable without requiring judges to create test data.

What we learned

Honestly, I learned about Medo on this hackerthon and Baidu's efforts to foster a relationship with developers. As for my project, I learned how serious the problem is globally, and it is shocking coz it is only growing.

What's next for Redcart

So the demo is definitely for the hackerthon, but the production-grade product will use the Django code generated by Medo, where I will host it externally, use a CDN for the screenshots, a self-hosted db, more security features (blurring personal details in the screenshots), adding admin moderation, analytics, adding social authentication, SEO optimization, and the hardest of all, Marketing and an awareness campaign. The sweetest part will be incorporating user feedback into the product.

Built With

• react
• superbase
• tailwind
• typescript
• vite