Alpha: automation for Hermes Agent

A-share stock screening system combining the Kronos K-line model with a Hermes Agent self-evolution loop

What Alpha does

Alpha is an open-source quantitative stock screening system for the Chinese A-share market that pairs the Kronos financial K-line foundation model with a Hermes Agent loop. The system observes the market, reasons about themes, validates signals and improves over time. Its screening works as a three-pool funnel, with intraday scoring and promotion and demotion rules for candidate stocks.

Hermes Agent connects through an MCP server over stdio that exposes 20+ tools, and handles scheduled post-market review, announcement review and live monitoring. Kronos forecasts appear in the research UI and APIs, paper trading simulates execution without live capital, and a FastAPI backend provides 50+ REST endpoints, WebSocket snapshots and an 8-tab dark dashboard. Data comes from sources such as Tushare Pro, Eastmoney, Sina, Tonghuashun and AkShare, and notifications can go to Feishu through a webhook. It runs with Docker or from source.

Key features

Three-pool funnel screening with intraday scoring
Kronos K-line forecast inference in the UI and APIs
Hermes Agent integration through an MCP server with 20+ tools
Paper trading for strategy validation without live capital
FastAPI backend with 50+ REST endpoints and WebSocket snapshots
8-tab dark web dashboard

When to use it

Run scheduled post-market and announcement reviews with a Hermes agent
Screen A-share candidates and validate strategies with paper trading
Fork the pipeline to build your own LLM-agent research workflow for Chinese equities

Who it is for: Researchers and builders working on A-share market analysis who want an LLM-agent screening workflow they can audit and modify.

How it fits with Hermes Agent

Built around Hermes Agent as the self-evolution loop, which drives the system's reviews and monitoring through its MCP tools.

How to install Alpha

These commands are copied from the project's README. Check the repository for the latest steps before you run them.

cp .env.example .env
docker compose up -d

Requirements: Docker (recommended) or Python 3 with a virtual environment; Tushare and LLM keys are optional

Note: It focuses on the Chinese A-share market, and its trading feature is paper trading with simulated execution.

FAQ

What is Alpha?

Alpha is an open-source, self-evolving quantitative stock screening system for the Chinese A-share market. It combines the Kronos K-line foundation model with a Hermes Agent loop.

Does Alpha work with Hermes Agent?

Yes. Hermes Agent is part of the architecture and talks to the system over MCP, handling scheduled post-market review, announcement review and live monitoring.

How do I install Alpha?

The recommended path is Docker: copy .env.example to .env, then run docker compose up -d and open http://127.0.0.1:18888. You can also run it from source with a Python virtual environment and ./start.sh.