InsightPotro

InsightPotro auto-generates slides from trends—turning weekly insights into shareable stories.

Comment

Inspiration

InsightPotro was born out of the frustration of sifting through raw data, trends, and charts every week without a clear summary or visual presentation. As students and builders, we often wished we had an assistant that could transform noisy data into clean, insightful, and presentable summaries — autonomously. With the power of multi-agent systems and Google Cloud, we realized we could automate the entire journey from data to deck.

What it does

InsightPotro is an autonomous, multi-agent system that: Collects real-time trend data on topics and locations Cleans, analyzes, and interprets the data Generates human-friendly summaries with key insights Creates a Google Slides presentation with: 3-part structured summary Visual chart of topic popularity Outputs a ready-to-share presentation link All of this happens with a single click via a modern frontend.

How we built it

Frontend: React + Tailwind CSS with Framer Motion for modern UI/UX Backend: Python Flask API

Multi-Agent System: CleanerAgent: Cleans raw data AnalyzerAgent: Extracts top insights and scores InsightWriterAgent: Writes detailed summaries PresenterAgent: Builds Google Slides using Google Slides & Drive API Charting: Matplotlib for generating visuals Storage: Local filesystem (can be upgraded to Firebase or GCS)

Challenges we ran into

Managing Google Slides API permissions dynamically Ensuring proper formatting of summaries across 3 slides Avoiding overlap between text and charts Handling delays from Google Drive image syncing Making the UI look clean while preserving accessibility

Accomplishments that we're proud of

Fully working multi-agent system from scratch Seamless integration with Google Slides End-to-end automation: from user prompt to shareable slide deck A modern, responsive UI that feels intuitive Robust error handling and modular agent-based design

What we learned

Real-world use of Google Slides & Drive APIs Building modular multi-agent systems Handling async workflows and API rate limits Designing UX that explains an AI system's output in a clear way Working with Framer Motion & Tailwind for production-ready frontend

What's next for InsightPotro

Add Firebase/GCP for storing user history and insights Enable voice command input ("Tell me what's trending in the US") Connect to X/Twitter, Reddit, or Google Trends as live data sources Add more slide templates (quote + stat, insight + chart, etc.) Launch a Chrome extension: "Turn this webpage into a slide"

Built With

Share this project:

Updates