secedgar

Diff every SEC filing against the last one, and see what companies quietly rewrote.

AI-assisted analyst desk that turns a 10-K’s reported earnings into economic earnings: AI extracts and classifies footnote items, code computes NOPAT, ROIC vs WACC and a rule-based rating, and the analyst validates every step.

End-to-end SEC XBRL financial data pipeline for Northern Trust (NTRS) — Python, Pandas, SQLite, SciPy | Data cleaning, inferential statistics on 1,200+ records

Tracks institutional 13F holdings from SEC EDGAR — SQL-driven quarter-over-quarter analysis with a Streamlit dashboard

FP&A Intelligence Dashboard for American Express - ML, ARIMA forecasting, SEC EDGAR data, Streamlit

Full-stack RAG platform that ingests SEC 10-K filings via EDGAR into a pgvector store, then answers natural-language questions about them using a retrieval pipeline backed by a local or remote LLM — with a Kafka-based ingestion queue, Prometheus/Grafana observability, and a Helm chart for Kubernetes deployment.

Detect edits and changes between SEC 10-K filings with a signal score and word-by-word markup.

Agentic RAG over real SEC 10-K filings — generates risk memos and independently verifies management's narrative against real XBRL financials, with a deterministic grounding check (not just LLM confidence). Live analysis for any ticker. FastAPI + Groq + ChromaDB/TF-IDF, deployed free on Render + Vercel.

RAG assistant for SEC 10-K filings with cited answers and groundedness evaluation