Introducing ValueCell: An Open‑Source Multi‑Agent AI Platform for A‑Shares, US Stocks and Crypto

Overview

ValueCell is an open‑source Python platform that uses a multi‑agent architecture to improve the efficiency of financial analysis and investment‑management tasks. It models a professional investment team, assigning distinct AI agents to responsibilities such as market analysis, sentiment analysis, and portfolio recommendation.

Core Features

Multi‑Agent System : separate AI agents specialize in different analysis domains.

Flexible Integration : supports multiple LLM providers, including OpenRouter and OpenAI.

Broad Market Coverage : includes US equities, A‑shares, Hong Kong stocks, and cryptocurrencies.

Community‑Driven : development and feature enhancements are contributed by a global community.

One‑Click Deployment : quick‑start scripts enable rapid setup.

Financial Focus : designed specifically for portfolio management and investment decision‑making.

Architecture and Workflow

The system is coordinated by an Orchestrator (the “conductor”). The Orchestrator incorporates planning, memory, storage, task‑management, and dialogue capabilities. When a user submits a query, the Orchestrator dispatches requests to Expert Agents via an A2A (agent‑to‑agent) protocol, aggregates their responses, and returns the final result to the user.

Alpha Arena AI Trading Competition Replication

ValueCell adds an AutoTradeAgent to reproduce the Alpha Arena competition, in which six top AI models each receive $10,000 of real capital to trade freely in cryptocurrency markets. The competition results show that only the domestic models Qwen and DeepSeek remain profitable, while the other models lose most of their capital; ChatGPT, for example, ends with $2,800.

ValueCell can generate line‑charts that plot each model’s equity curve on a shared timeline, allowing direct visual comparison of performance across market conditions.

Deployment Guide

Follow these steps to run ValueCell locally:

Clone the repository:

git clone https://github.com/ValueCell-ai/valuecell.git

Enter the project directory: cd valuecell Copy the example environment file and edit it with your API keys (e.g., OpenRouter): cp .env.example .env Modify .env accordingly.

Start the application:

Linux/macOS: bash start.sh Windows: .\start.ps1 Open a browser and navigate to http://localhost:1420 to access the UI.

Log files are written to the logs directory for troubleshooting.

Docker deployment is also supported; refer to the repository for Docker instructions.

Repository: https://github.com/ValueCell-ai/valuecell