BiasBuster - Cut through the spin. Spot the bias.

• 

  Next Results

• 

  First Results

• 

  Initial Screen - Analyzing

Inspiration

I’ve always been fascinated by how language shapes perception. From marketing copy to political speeches, subtle cognitive biases influence how we interpret information—often without realizing it. I wanted to build a tool that helps people see through persuasive tactics and think more critically about the content they consume.

What it does

BiasBuster is a Chrome extension that analyzes online content using on-device AI to detect cognitive biases. It breaks down paragraphs in real time and highlights biases like Authority Bias, Halo Effect, Bandwagon Effect, and more—giving users a clearer lens on the messaging they’re reading.

How we built it

I built BiasBuster using Chrome’s Manifest V3 framework and the built-in LanguageModel API (Gemini Nano). The extension extracts article text, splits it into paragraphs, and sends each one to the model in parallel. I designed a clean, responsive popup UI that updates in real time as results come in.

Challenges we ran into

The biggest challenge was performance. Sending large blocks of text to the model caused delays or silent failures. I had to experiment with input cleaning, chunking strategies, and prompt tuning to make the experience fast and reliable. UI responsiveness was also tricky—especially syncing loading states with asynchronous model calls.

Accomplishments that we're proud of

I’m proud that BiasBuster runs entirely on-device, preserving user privacy while delivering meaningful insights. The real-time paragraph analysis feels smooth and intuitive, and the bias detection results are surprisingly accurate. Seeing it work on live websites was a huge moment.

What we learned

I learned how to work with Chrome’s AI APIs, optimize prompt design for lightweight models, and build a user-friendly extension that balances speed with depth. I also gained a deeper appreciation for how subtle language cues can shape beliefs—and how tech can help us challenge them.

What's next for BiasBuster - Cut through the spin. Spot the bias.

Next, I want to:

• Add bias icons and severity indicators
• Support multilingual analysis
• Let users export results or share insights
• Build a bias glossary for educational value
• Explore auto-analysis on page load

BiasBuster is just getting started—and I’m excited to keep pushing it forward.

Built With

• chrome-built-in-ai-api-(languagemodel)
• chrome-extensions-(manifest-v3)
• chrome-scripting-api
• css
• gemini-nano
• html
• javascript