Inspiration
The web holds enormous amounts of valuable information, yet extracting and organizing that data efficiently remains a challenge. I wanted to create a tool that simplifies this process — something that could intelligently extract structured data without violating privacy or terms of service. The idea was to design a system that combines the reliability of local execution with the power of cloud-based AI.
What It Does
Web Weaver Dual Mode is a Chrome extension that extracts and structures website content into JSON or CSV formats. It offers two modes:
Chrome Built-in AI for fast, on-device, privacy-preserving extraction.
Gemini Cloud API for advanced tasks such as content classification, data pattern recognition, and contextual understanding.
This dual approach ensures flexibility — users can prioritize privacy or performance depending on their needs.
How It Was Built
The extension is built using React and Tailwind for the frontend, integrated with Chrome’s Manifest V3 framework. For cloud operations, it connects to the Gemini API to process and structure extracted data. The system is modular, allowing seamless switching between local and cloud modes, with Chrome’s local storage managing session data and user preferences.
What I Learned
This project deepened my understanding of how to balance speed, privacy, and scalability in AI-assisted systems. I learned to integrate Chrome’s experimental AI capabilities effectively and handle asynchronous data extraction while maintaining compliance with web content terms of service.
Challenges
One of the primary challenges was designing an extraction method that respects site policies while remaining accurate and efficient. Handling dynamically loaded web pages and maintaining consistency between the two AI modes required extensive optimization and testing. Another challenge was ensuring stable integration with Gemini APIs, which often involved managing latency and response parsing.
Future Work
Future versions will include voice-based extraction commands, richer AI analysis modes, and direct export options to third-party tools like Google Sheets. The long-term goal is to make Web Weaver an open, extensible platform for AI-powered web data interaction.
Log in or sign up for Devpost to join the conversation.