Recommi

Recommi is an app that uses AI to rank and recommend anything to you summarised in a simple list. Just tell Recommi what you're looking for and the app will find the best recommendations for you.

Comment

Inspiration

Throughout life, we constantly seek recommendations — from family, friends, and search engines — whether it’s for products, places, or experiences. But the process of finding a trusted recommendation is messy. There are scattered conversations, endless group chats, conflicting opinions, overwhelming ad-filled search results. There’s no simple, organized, or reliable way to gather, compare, and act on these recommendations easily. There has to be a better way — a simpler, smarter way to find and manage trusted recommendations all in one place.

What it does

Recommi finds and ranks recommendations about anything you ask and presents it in a simple, beautiful list. You can:

  • Search for recommendations about anything
  • Read concise summaries for each recommended item
  • See what people in the internet have said about each item
  • View images about each recommended item
  • View the location on a map for places recommended
  • Save the list of recommendations for later

How we built it

  • Used Perplexity’s Sonar API's web search to fetch recommendations and images and then provide the results ranked
  • Developed a mobile app using Flutter for cross-platform development
  • Designed with Figma, and animated with Rive
  • Used Firebase Cloud Functions as a serverless backend
  • Focused on creating a smooth mobile experience to counter latency of AI-generated results

Challenges we ran into

  • Maintaining consistency in AI response formats to ensure a clean, reliable user experience
  • Handling streaming responses properly and parsing partial outputs without errors
  • Trading off between quality, cost, and latency of AI calls to achieve the best possible experience
  • Optimizing for mobile UX where users expect fast, fluid interactions even when waiting for AI-generated results

Accomplishments that we're proud of

I’m proud to have designed and developed a full end-to-end product that solves a real need in my own daily life. Tackling technical challenges like AI integration, response parsing, and creating a seamless mobile UI was a major milestone. Building Recommi as a solo effort — handling everything from backend to frontend to AI orchestration. It was challenging, but it was a lot of fun creatively and technically.

What we learned

This project opened my eyes to the power of Perplexity’s Sonar API, particularly for real-time search-based AI applications. I also deepened my understanding of handling streaming responses, optimizing serverless functions for mobile apps, and designing user-centric AI experiences.

What's next for Recommi

  • Incorporate features to build user trust (e.g., citations, confidence scores, category experts)
  • Add a social element so users can share and comment on recommendation lists
  • Introduce personalization based on user preferences and past behavior
  • Tailor the recommendation experience based on the category (e.g., different UX for food, travel, or tech)

Ultimately, I want Recommi to be the app people go to for finding the best trusted recommendations

Built With

Share this project:

Updates

Submission history