stocks-insights-ai-agent

Why we included this project

Most write-ups about agentic RAG stay at the diagram level; this repo is the working version, applied to stock data. It pairs LangChain and LangGraph with ChromaDB so you can ask plain-language questions about a company's price history and news, and the answer comes back with charts. The news workflow is the interesting part: it retrieves from the vector store, scores how relevant the documents are, and falls back to a live web search when nothing local fits, which is exactly the loop that separates a demo from something usable. Financial data gets scraped into PostgreSQL and queried through generated SQL, and the project ships an OpenAPI spec plus pytest coverage, so it works as a template for structuring a full-stack LLM service. It is a single-contributor demo, so treat it as a blueprint to adapt, not a product to deploy as-is.