Inspiration
Shopping online can be overwhelming with so many options and unclear product descriptions. We wanted to create a smart shopping assistant that helps users quickly gather insights about products, compare alternatives, and make informed decisions that are all within a seamless Chrome extension.
What it does
clerk is a Chrome extension that:
- Reads the URL of the product you're viewing
- Scrapes product details and similar recommendations
- Displays LLM-powered insights based on product descriptions
How we built it
Frontend: Built using TypeScript, JavaScript, and HTML/CSS, the extension interacts with the browser’s active tab, fetches URLs, and displays chatbot responses in a sleek UI. Backend: Manages user queries, stores messages, reads product data, and returns AI-generated responses from a custom-built chatbot using llama. All through file based communication Webscraping-Extracts product details, specifications, and user reviews from e-commerce websites.
Challenges we ran into
- Initially, the extension required manual activation. We had to implement background scripts to capture and update URLs automatically.
- The chatbot’s responses didn’t appear instantly, so we had to adjust file monitoring and reloading mechanisms.
- We had to rotate user-agents and introduce delays to avoid getting blocked.
Accomplishments that we're proud of
- Successfully built a seamless chatbot experience inside a Chrome extension
- Integrated real-time web scraping to fetch product details
- Overcame CORS and URL tracking issues to ensure smooth functionality
- Designed a minimal and stylish UI that enhances the user experience
What we learned
- How to integrate background scripts for continuous webpage tracking
- The importance of handling CORS & security permissions in Chrome extensions
- Web scraping best practices and avoiding detection on e-commerce sites
- How to build file-based communication between a browser extension and a backend
What's next for clerk
- Automatically apply filters based on user-provided search parameters.
- Price Tracking: Add alerts for price drops and historical trends.
- Expand to more websites
- Mobile App Version: Bring clerk’s shopping assistant to iOS & Android
- Improve UI
Built With
- beautiful-soup
- flask
- javascript
- langchain
- llama
- milvus
- python