The tool ingests multi‑page financial reports, parses them with IBM Docling, chunks and stores embeddings in ChromaDB, retrieves relevant sections via Cohere reranking, and uses Gemini 1.5 Pro to extract and validate over 50 financial fields into JSON. It targets finance analysts and data scientists needing automated, high‑accuracy extraction from PDFs.
View on GitHub →Enxt-AI/MCA