Terms Risk Scanner

Inspiration

I’ve noticed that most people (including me) don’t actually read terms and policies. Not because we don’t care, but because they’re long, scattered, and hard to understand. Important details usually show up after you’ve already clicked “Agree” or “Continue.” I wanted to build something that helps people understand what they’re signing up for before making a decision.

What it does

Terms Risk Scanner is a browser extension that scans the current webpage and highlights parts of the terms or policies that could affect the user. It pulls out things like:

• auto-renewals
• cancellation rules
• refunds
• data usage or sharing language It then explains these findings in simple language so users can make a more informed choice without digging through multiple pages.

How I built it

The extension is built using Browser Use Skills. I use browser automation to:

• navigate real webpages
• read publicly available terms and policy sections
• extract relevant information in a structured way The extension displays the results directly in the browser, keeping everything lightweight and easy to access.

Challenges I ran into

One major challenge was that many signup or free-trial flows require creating an account or logging in, which isn’t reliable to automate. To work around this, I focused on pages that are accessible without login, such as:

• terms pages
• pricing pages
• help and policy pages Another challenge was deciding what information actually matters to users, and keeping the output short and understandable instead of overwhelming.

Accomplishments that I’m proud of

• Built a working browser extension that integrates real Browser Use Skills
• Turned long, complex terms into something users can quickly understand
• Designed the extension to feel like a real product, not just a demo
• Kept the solution practical and safe to run on live websites

What I learned

I learned how to code and work much closer to the technical side than I had before. Building this as a solo project pushed me to understand how browser automation, APIs, and extension logic actually work, not just how they should work in theory. I also learned that the hardest part isn’t extracting information, it’s deciding what’s worth showing to the user. Turning raw data into something clear, useful, and not overwhelming required a lot of iteration.

What’s next for Terms Risk Scanner

Next, I’d like to:

• improve how I summarize and group findings
• add comparisons to common or “standard” terms
• expand support to more types of websites

The long-term goal is to help users catch important details early, before small decisions turn into long-term regret.

Built With

• browser-use
• codex
• vscode