Developer launches self-hosted MCP server Equibles to feed local LLMs real-time financial data

One of the key limitations when running large language models locally as autonomous agents is their inability to access current, real-world data. To solve this, developer DanielAPO created Equibles—a self-hosted, open-source MCP (Model Context Protocol) server that scrapes and serves public U.S. financial data directly to any MCP-capable client. Equibles operates entirely on your machine, requiring no cloud services, no API keys, and no telemetry. It works out of the box with tools like Claude Code, Claude Desktop, Cursor, or custom local-model agent loops, enabling them to query financial datasets instantly.

Equibles provides a comprehensive set of financial data sources: full-text search across SEC filings (10-K, 10-Q, 8-K), institutional holdings from 13F filings, insider trades (Form 3/4), congressional trades, FINRA short volume and short interest data, SEC fails-to-deliver, FRED economic indicators, CFTC futures positioning, and CBOE VIX/put-call ratios. It also serves daily prices and technical indicators. The project is hosted on GitHub at github.com/daniel3303/Equibles, where the developer is actively seeking feedback and feature suggestions. For privacy-conscious professionals and developers, Equibles offers a powerful way to integrate real financial intelligence into local AI workflows without compromising on data sovereignty.