MCP Stock Terminal

Overview

A local AI financial research terminal. Allows conversational analysis of stock market data using @ticker mentions and /commands, correlating stock price volatility with market news to explain why a security moved.

How it works

1. 01

   Native Model Context Protocol (MCP) client-server architecture with domain logic partitioned into discrete tools.

2. 02

   Local LLM inference using Ollama running qwen2.5:14b.

3. 03

   Vector embeddings and semantic financial news search backed by pgvector in PostgreSQL.

4. 04

   Contextual news-to-price correlation explaining catalysts behind historical intraday movements.