What it does

ClearFrame is an intelligent platform that analyzes text and images to uncover deeper meaning.

Detects emotion and sentiment behind text Identifies biases (political, social, cognitive) Scores manipulation level Generates neutral, fact-based rewrites Extracts text from images using OCR Works directly inside the browser via extension

It basically helps users go from “reading content” → “understanding content deeply”.

How we built it

I built ClearFrame using a modern full-stack approach:

Frontend: Next.js 15, Tailwind CSS, Radix UI Backend: Inngest (for background jobs & workflows), MongoDB AI: Groq SDK + Google Gemini Auth: Clerk Other tools: Tesseract.js (OCR), GSAP (animations)

Inngest plays a key role — it handles long-running AI tasks and workflows, which is important because AI processing can’t always run instantly in serverless environments.

Challenges we ran into

This project wasn’t easy — especially on the backend side:

Very slow responses due to heavy AI processing Token limits getting exhausted quickly Vercel limitations (can’t handle long background jobs properly) Managing async workflows and retries Balancing performance vs cost

These were real bottlenecks, especially when trying to make the app feel fast and responsive.

Accomplishments that we're proud of

Built a fully working AI-powered analysis platform Integrated multiple AI providers smoothly Created a browser extension + web dashboard sync Designed a UI that feels modern and smooth The system is fast, portable, flexible, and relatively lightweight

For a project like this, getting everything working together is a big win.

What we learned

This project taught me a lot:

Real-world AI integration is complex Managing tokens and cost optimization is critical Basics of RAG (Retrieval-Augmented Generation) How to design better UI/UX for AI tools Importance of animations (GSAP) for user experience How to structure serverless workflows using Inngest

It was a mix of experimentation and learning by doing.

What's next for ClearFrame

There’s still a lot to improve:

1 Make responses faster and more efficient 2 Improve AI understanding and accuracy 3 Better authentication + user personalization 4 Stronger browser extension integration for instant results 5 Add contextual tips/insights for users 6 Deeper integration with Next.js for smoother workflows 7 Optimize token usage and backend performance

The goal is to make ClearFrame feel instant, intelligent, and effortless to us