Agent 10-K

Talk to Your Reports. Track the Real Story. An AI agent that reads annual reports, answers your questions, builds slide decks, and fetches real-time company news — all from a single PDF upload.

Comment

Inspiration

We wanted to make dense financial documents like annual reports more accessible and actionable. Reading 200+ pages shouldn't be a prerequisite to understanding a company.

What it does

Agent 10-K reads uploaded PDFs, answers natural language questions, summarizes insights into PowerPoint slides, and fetches live company news using SerpAPI.

How we built it

We used LangChain agents with Gemini Pro and Google Embeddings to build a PDF-based RAG system, integrating SerpAPI for real-time news and Streamlit for the front end.

Challenges we ran into

Handling unstructured PDF data was tough, especially preserving context. Integrating agent tools while maintaining response accuracy also required careful orchestration.

Accomplishments that we're proud of

We built a seamless, multi-tool AI agent that not only understands documents but augments them with real-time web knowledge — all with a user-friendly interface.

What we learned

We learned how to effectively chain LLM tools, manage vector stores, and blend static and dynamic content for enterprise-grade document analysis.

What's next for Agent 10-K

We plan to add financial chart extraction, multi-PDF comparison, voice queries, and memory support for deeper, more interactive document analysis.

Built With

Share this project:

Updates