FeedlyAI

Inspiration

Social media can be overwhelming — endless feeds, long threads, and the pressure to post the “right” content. We wanted to create a smarter, AI-powered assistant that helps users discover, create, and interact with content effortlessly, without feeling lost in the noise.

What it does

FeedlyAI is an AI-driven social media assistant that suggests posts, captions, and hashtags based on a user’s mood or interests. It can summarize long threads or trending posts into quick, digestible insights and recommend thoughtful or humorous replies to conversations. Additionally, it filters the feed by tone or mood, allowing users to focus on content that is inspiring, funny, or relaxing. Overall, it creates a smarter, faster, and more personalized social media experience.

How we built it

We built FeedlyAI using React and Tailwind for a smooth, minimal frontend interface. The backend runs on Firebase, which handles authentication, real-time database interactions, and hosting. The AI functionality is powered by the OpenAI GPT API, which generates content suggestions, summarizes threads, and recommends replies. We also incorporated sentiment analysis to support mood-based feed filtering and added basic engagement predictions to enhance the user experience. Our focus was on delivering a hackathon-ready MVP with a functional feed, AI suggestions, and a seamless posting workflow.

Challenges we ran into

One of the biggest challenges was teaching the AI to generate contextually appropriate posts that did not feel repetitive. We also had to handle multiple users interacting in real time while keeping the feed responsive and smooth. Another challenge was balancing AI suggestions with user control, ensuring that the AI enhanced the experience without taking over the feed entirely.

Accomplishments that we're proud of

We successfully built a fully functional AI-powered social feed in just a few days, complete with smart post suggestions that felt natural and engaging. Our demo clearly showed how AI can enhance social media interactions, making posts easier to create, threads easier to understand, and conversations more meaningful.

What we learned

Through this project, we learned that AI can make social media more manageable and enjoyable, and that small features, like summarizing threads, can have a big impact on user experience. We also realized that hackathon MVPs work best when the team focuses on one core AI feature rather than trying to build everything at once.

What's next for FeedlyAI

Moving forward, we plan to add multi-platform posting so that users can share AI-suggested content across Instagram, X, TikTok, and more. We aim to implement personalized learning so that the AI adapts to each user’s style over time. We also want to explore AI-driven conversation moderation to improve online interactions and test scalable deployment to support a large user base.

Built With

• firebase
• hugging-face-api
• javascript
• node.js
• openai-gpt-api
• python
• react
• tailwind-css