Inspiration

As a beginning-level investor, I often struggle to judge whether the information I come across actually influences the stock prices I hold, especially the news and financial articles that vary quickly even in a day. This personal challenge inspired me to create InsightBridge, a simple real-time AI-driven Chrome extension that helps investors instantly identify whether a webpage’s content is relevant to their portfolio.

What it does

InsightBridge analyzes the webpage you’re currently viewing and provides:

  • A clear indication of relevance.
  • Bullet-point explanations supporting the relevance conclusion.
  • A concise five-sentence summary of the key content for better context understanding.

The app also features cache memory, allowing it to remember the last inputs. Users can continue where they left off without re-entering information of the concerned stock and language preference.

How I built it

I developed InsightBridge using:

  • Chrome Extension API for webpage data extraction, sidebar interface, and background event management.
  • Google AI Prompt API and Summarization API to perform content understanding, summarization, and relevance reasoning.
  • Yahoo API as an external data source to validate and identify stock tickers and company information.
  • HTML, JSON, and JavaScript for the browser interface and logical flow between front-end display and back-end AI functions.

The system architecture integrates multiple APIs in a modular way: the workflow starts from retrieving the company with the provided stock sticker via Yahoo API. Then collecting the webpage information by background script and summarizing via Summarizer API. Next, the Prompt API analyzes the relevance and gives a wrap-up summary. The content is well-formatted before presenting on the side panel.

Challenges I ran into

The main challenge was my limited background in JavaScript and HTML, which are essential for building Chrome extensions. Learning asynchronous logic, DOM manipulation, and cross-context communication required significant time and experimentation. The outlook design of the app is also difficult to balance simplicity, design aesthetics, technological sophistication, and professionalism.

Accomplishments that I'm proud of

I’m proud of building a functional AI-powered Chrome extension entirely on my own — learning full-stack web development, API integration, and AI orchestration along the way. It is also an acomplishment that the stock ticker identification process is stablized in current version. Transitioning from unstable Prompt API outputs to a hybrid data-driven system was a major improvement that made the app consistent across different web sources.

What I learned

I learned how to bridge browser-based UI logic with AI reasoning APIs in real time. Implementing cache memory taught me about persistent local storage and user data handling, while integrating Yahoo and Google APIs improved my understanding of how to balance latency, accuracy, and usability. Most importantly, I learned how to iteratively refine AI behavior for real-world use by testing, adjusting prompts, and improving data flow structure.

What's next for InsightBridge

Future improvements will focus on expanding the tool’s capability and inclusiveness:

  • Enhance cache efficiency to support multi-session use and faster load times.
  • Introduce multi-stock comparison, allowing users to see which holdings are most relevant to a given article.
  • Expand the stock database to include Hong Kong and Shenzhen-listed companies, making the tool useful for regional investors in Greater China.
  • Add more language options, especially Chinese (Simplified and Traditional), to ensure accurate analysis across multilingual financial content.
  • Develop an aggregated insight dashboard summarizing daily relevant news per stock and tracking emerging trends.
  • Evaluate the fidelity of the webpage information.
  • Improve UI design and summarization readability across both desktop and mobile devices.

My long-term goal is to make InsightBridge a trusted personal assistant for investors — transforming information overload into clear, AI-driven investment insight.

Built With

Share this project:

Updates