DocZilla

• 

  Team members and our contacts

• 

  Logic & Architecture of Doczilla

• 

  Problems Doczilla tackling

Inspiration

After mentorship sessions and insightful conversations with representatives from UpstageAI, we identified a major challenge that companies face: spending billions annually and countless hours on manual document processing, contract reviews, and compliance checks. Witnessing firsthand how much valuable productivity is lost inspired us to leverage UpstageAI’s powerful API to create DocZilla—an intelligent AGI-powered document assistant using a RAG pipeline to automate these tedious tasks.

What it does

DocZilla automatically analyzes contracts, agreements, and policy documents against internal company standards. It instantly flags compliance issues, identifies risky clauses, extracts key action items and dates, and seamlessly integrates with everyday productivity tools like Slack, Notion, and Google Calendar. It saves time, reduces human error, and ensures no important detail is overlooked.

How we built it

We built DocZilla using a React and Next.js frontend paired with a robust FastAPI backend. Leveraging Solar LLM from Upstage, we implemented Retrieval-Augmented Generation (RAG) to quickly and effectively analyze uploaded documents. For vectorized database we used FAISS, for integrations, we connected APIs from Slack, Notion, and Google Calendar to automate workflow actions, document sharing, and task management.

Challenges we ran into

One major challenge was ensuring accurate and rapid document analysis without overwhelming the user with too much information. Also having only 2 members in a team was definitely challenging, but rewarding at same time. Integrating the AGI backend with intuitive UI interactions required a lot of testing and refinement.

Accomplishments that we're proud of

We successfully created an intuitive and streamlined user experience that makes complex AGI features accessible to everyone. Our seamless integration with popular productivity tools significantly boosts workflow efficiency, and we’re especially proud of our effective use of RAG to improve document analysis speed and accuracy, connected with smart actions creating optimized workflows

What we learned

We learned a lot about combining powerful AGI models with user-friendly design principles. Working through technical integration challenges deepened our skills in full-stack development, API integrations, and AI-enhanced workflows. We also learned the importance of rapid iteration based on user and mentor feedback.

What's next for DocZilla

We plan to enhance and scale DocZilla’s capabilities with advanced features like automated compliance audits, predictive risk assessments, and deeper and more integrations with more platforms. We also want to work on scaling our RAG solution, to speed up processes. Our goal is to continuously refine DocZilla to become the go-to intelligent assistant for effortless and accurate document management.

Built With

• faiss
• fastapi
• langchain
• nextjs
• openai
• react
• typescript
• upstageapi