After tinkering for a while, I finally turned an idea into an open-source project — ValueCell: the world’s first open-source platform for financial agents The core idea is simple: take the scattered modules in the investment research workflow (market research, data scraping, sentiment analysis, strategy generation, risk monitoring…) and turn each into its own Agent. Then connect them together into a collaborative, extensible research ecosystem. How it works Each Agent is like a “cell” — focused on doing one thing well (say, sentiment analysis). They can be freely combined, cross-check each other, and ultimately generate richer investment insights. Since it’s open-source, anyone can build or improve Agents and help grow the ecosystem.

We built ValueCell to be open and compatible. It works seamlessly with mainstream frameworks like Agno and LangChain, integrates at the agent level with TradingAgents and AI-hedge-fund, and connects to trusted data sources such as Yahoo Finance and AKShare. On the technical side, we also added support for MCP and A2A, ensuring smooth interoperability across agents and tools.

What it solves Traditional research: fragmented tools, low efficiency, info overload. ValueCell: multi-agent collaboration → turning scattered data & opinions into actionable insights.

This is just the beginning — there’s still a lot to refine. But I want it to be an open playground for research & experimentation. Whether you’re a dev, a researcher, an investor, or just curious about AI + investing, you’re welcome to jump in. Give it a on GitHub, and if you’re interested, let’s build together!