Ripple

Matching up and coming brands with high-performing microinfluencers using AI

Comment

Inspiration

The inspiration of Ripple all started with just another day on YouTube. I came across a video of 15-year-old Evan Yadegari, an app founder who has been quite successful in the space. He was talking about all the challenges he faced and what ended up working out. Surprisingly to me, he said that by far, the hardest part was handling influencer partnerships, where he spent hours and hours scouring social media for creators that fit, censuring nothing is botted, looking for high comment and like rates, deciding what type deal to offer (flat fee, CPM, or both), and making personalized messages to each one. The worst part was that 95% of these influencers he put so much time into finding never even bothered to respond. In my opinion this issue is such a massive waste of time for founders that prevents them from doing what they truly love: building, causing me to think of an AI solution for this, which was Ripple.

What it does

Ripple is a platform for early-stage startups and e-commerce brands where all you have to do is scan your site, and it gives you 20-30 perfectly matched social media creators from TikTok, Instagram, and YouTube that could promote your business by weaving it into the creator's natural content. It gives you a overall compatibility score, a suggested personalized email or DM you could send to them, and a way to have the creator weave the ad into their content.

How we built it

Ripple is built with TanStack Start, a full stack React framework with SSR/server functions and Tailwind CSS v4. It uses API calls to YouTube, TikTok, and Instagram to find creators, in addition to Firecrawl and a Gemini API for site scanning, matching, and personalized outreach DMs. The frontend runs as a server-rendered React app, with backend logic handled by createServerFn server functions rather than a separate API server.

Challenges we ran into

While building Ripple, we ran into problems, just like all projects. One of the hardest of those problems was trying to get Instagram and TikTok data for each creator, especially considering how it's harder to acquire API keys for these platforms and that crawlers were necessitated. Another one is...

Accomplishments that we're proud of

I'm very proud of the whole process of building Ripple. I'm proud especially that my brother and I made it together, a testament to our shared interests and I'm really happy we got great bonding time. Another thing I'm proud of is also a result of our sibling status: that since we are always together and were working hard, we made the whole project in under 5 days. The last thing I would say I'm proud of is that when we make apps in the future, we can just use Ripple to help scale it if it reaches that level

What we learned

I learned a lot in the experience of building Ripple, which was my 2nd full stack platform. One of many lessons I learned from this experience is teamwork since it was made by my brother and I. We would have disagreements on various aspects of the app like UX or the backend but we always managed to figure out a solution we agreed on. Moreover, I learned a lot about designing a platform in terms of UI. Having a nice design scheme is key when building a platform or app, especially incorporating newer UI ideas like 60-30-10 splits and glassmorphism. I've learned that even if your app is the best thing in the world, nobody's going to want to use it if it looks horrible.

What's next for Ripple

It's hard to predict what's next for Ripple but I hope to make some promotional videos on YouTube. We need to get people on the site to gather initial feedback and improve the product further. We will also use it ourselves, for our apps we make in the future that will hopefully be successful.

Built With

Share this project:

Updates

Submission history