The integration of artificial intelligence with financial markets has reached a fundamentally new level. An analyst under the pseudonym CyrilXBT demonstrated how, with a single command, Claude Code can be connected to live market data covering more than 17,000 stocks and cryptocurrencies. This involves access to quotes, earnings reports, balance sheets, and cash flow statements — functionality previously available only through the Bloomberg Terminal.
According to the expert, the setup takes about 60 seconds and consists of three simple steps. The first is running a command inside Claude Code to add the external service financial-datasets via HTTP transport using the MCP protocol. The second is authorization via OAuth in the browser. The third is natural language queries: for example, you can ask about the current price-to-earnings ratio and market capitalization of Apple, Tesla's quarterly results for the last four periods, or the change in Bitcoin's price over the year.
Democratization of analytics
The key point here is cost. The Bloomberg Terminal costs analysts approximately $24,000 per year. This kind of AI integration with live data effectively replaces that expensive tool. According to CyrilXBT, a sustainable advantage will go to those who learn to combine Claude Code's reasoning with access to up-to-date financial information — quants, analysts, and portfolio managers.
It is important to emphasize that each step individually is not difficult. The value arises precisely from combining an ordinary tool with a stream of live market data in a single interface. This is not just convenience but a paradigm shift: the open MCP standard allows AI to act as a shell through which the user accesses a third-party data provider without manually gathering information.
Scaling agents: from smartphone to thousands of processes
Also worthy of attention is another publication by CyrilXBT dedicated to the workflow of Boris Cherny, one of the creators of Claude Code. Cherny uses a smartphone as his primary interface instead of a laptop, while simultaneously maintaining five to ten active sessions. Each of them is capable of spawning auxiliary agents — sometimes several hundred, and sometimes several thousand overnight for larger tasks.
The expert highlights four elements without which such a scheme is risky: a full-fledged interface for sending instructions from a smartphone, isolated parallel execution, verification without human involvement, and strict stopping conditions with a spending cap. It is precisely this that turns overnight agent work from a risky venture into a manageable task.
My view: the integration of AI with live market data is not just a reduction in the cost of accessing information, but a fundamental shift in how investment decisions are made. However, it is worth remembering: the speed of data access does not replace the depth of analysis. The tool becomes dangerous if used without a clear methodology and risk control. Those who master this stack first will gain a colossal advantage, but only if they maintain discipline in validating results.