StockDocs — MCP

Data API server for the StockDocs financial news platform: exposes the processed article corpus and vector database over HTTP so LLM clients (via MCP wrappers) can query semantic search results, article content, and company facts.

Part of the StockDocs project family:

Repo	What it is
StockDocs	Processing core — article server, NLP analysis, embeddings
stockdocs-scraper	RSS scraper — collects the article corpus

Endpoints

Method	Path	Description
POST	/query	Semantic search over the ChromaDB vector store
POST	/articles/query	Filter the article corpus by criteria
GET	/articles/latest/<field>	Latest articles by field
GET	/company/<name>/facts	Company fact sheet
GET	/company/<name>/products	Company product list
POST	/company/facts/update	Update a company fact sheet
POST	/facts	Extracted-facts query
GET	/health	Health check
GET	/info	Service info
GET	/openapi.json	OpenAPI 3.0 spec

Running

pip install -r requirements.txt
python server.py    # :5005

Requires a running ChromaDB instance (host/port via environment).

Docker

docker build -t stockdocs-mcp .
docker run -p 5005:5005 -e CHROMADB_HOST=<host> stockdocs-mcp

Project Structure

.
├── server.py          # Flask app — all endpoints
├── openapi.json       # OpenAPI 3.0 spec (also served at /openapi.json)
├── requirements.txt   # Python dependencies
└── Dockerfile         # python:3.12-slim image