Clarity

Making the internet transparent, one document at a time

Inspiration

Clicking "I agree" when we haven't read the Terms of Service and Privacy Policy costs more in the future than the time we think it saves.

Earlier this year, my Wise account was suddenly banned for violating their Terms of Service, "which I never actually read." As a global remote worker, this wasn’t just inconvenient; it froze my main source of income and created weeks of unnecessary stress. (Thankfully, they allowed me to withdraw my balance later.)

A friend had a similar experience. She woke up to the message: "Access to your account has been suspended due to violation of our terms of service" on GitHub, and she didn't know what term she violated. sample message on X

And it’s not just account bans. Every day, we sign up for services whose Privacy Policies quietly mention that our personal data can be shared or even sold to third parties (advertisers). We skip through the fine print because it’s written in dense, legal jargon, and because we assume it won’t matter.

But it does matter. Our financial accounts, our identities, and our privacy are all bound by contracts we don’t understand, yet agree to daily.

What it does

Clarity helps people see what they’re really signing up for. Because informed consent on the internet shouldn’t require a law degree. They should know:

What actions will make their accounts suspended or banned.
Which of their personal information is collected.
How is their personal information collected.
How is their personal information stored, shared, or sold.

How we built it

We built Clarity with a shared mission: make the internet more transparent by helping users truly understand the Terms of Service and Privacy Policies they agree to every day.

Clarity was built primarily with TypeScript, leveraging React for the front-end and Node.js for the background and data processing scripts. We used Chrome Extension APIs to integrate directly with users’ browsing experience, so they can have Clarity of web legalese with just a right-click.

Our development process focused on:

Simplicity and performance: We designed Clarity’s popup and context menu for quick, intuitive access.

Scalability: The backend was built modularly, allowing us to plug in new document sources (e.g., Refund Policy, Cookie Policy) without breaking existing functionality.

AI integration: We used built-in Chrome AI capabilities (Prompt API and Translator) along with OpenAI’s models (as fallbacks) to parse dense legal texts, extract key points, and summarize them in plain English.

Data pipeline: We implemented automated web crawlers that retrieve, classify, and version key documents from websites, ensuring Clarity stays updated as companies revise their policies.

Challenges we ran into

Building Clarity came with its fair share of hurdles, both technical and procedural.

One of the biggest challenges was parsing HTML documents directly from websites. Due to CORS and Same-Origin Policy restrictions, it wasn’t possible to fetch and process these pages within the browser environment. We had to design creative workarounds, including using our backend as a middle layer to safely retrieve and sanitize the documents before analysis.

Another major challenge was working with the Built-in Summarizer API’s small context window, which wasn’t enough for long or complex Terms of Service documents. To overcome this, we switched to using the Prompt API, which gave us more flexibility and control over chunking and summarizing content efficiently without losing context.

Lastly, publishing the extension to the Chrome Web Store turned out to be more difficult than expected. The review and approval process required multiple adjustments to meet Chrome’s new policies on permissions, data handling, and manifest compliance, a learning curve that taught us the importance of transparency not just for users but for developers too.

Despite these challenges, each obstacle pushed us to refine Clarity into a more robust, compliant, and privacy-conscious tool.