A quiet but significant shift has occurred in the world of financial analytics. I have carefully studied the latest developments that are fundamentally changing the rules of the game for those accustomed to working with market data. Specifically, the AI tool Claude Code can now directly access live quotes for more than 17,000 stocks and cryptocurrencies, as well as financial reports and cash flow movements. In essence, this calls into question the need for expensive professional terminals.
How it works and what exactly is available
At its core, it involves connecting an external financial data service via the MCP protocol. The entire setup, as my analysis showed, takes no more than a minute and consists of three simple steps. First, a command is executed inside Claude Code to add the financial-datasets service via HTTP transport. Then, authorization is completed via OAuth in the browser, and finally, the ability to ask queries in natural language becomes available.
The practical value of this solution is enormous. Instead of manually gathering data from dozens of sources or writing complex API requests, an analyst can simply ask about Apple's current price-to-earnings ratio, Tesla's quarterly reports for the last four periods, or Bitcoin's price dynamics over a year. All of this is in a single interface, without switching between windows and without the need for a Bloomberg terminal, which costs users approximately $24,000 per year.
Who benefits from this integration
A sustainable advantage will go to those who can effectively combine AI's analytical capabilities with access to current market data. This includes quants, analysts, and portfolio managers. Moreover, this advantage accumulates daily, like a research edge that grows with each use. It is important to understand: none of the steps individually is complex, but it is precisely their synergy within a single tool that creates real value.
Scaling: thousands of agents from a smartphone
Another aspect deserves special attention—the ability to manage entire armies of AI agents. One of Claude Code's key developers, as I managed to find out, uses a smartphone as the primary interface, maintaining between five and ten active sessions simultaneously. Each of these can spawn hundreds or even thousands of auxiliary agents overnight to handle large-scale tasks. At the same time, dozens of cycles run continuously in the background, and some of them are saved on the server even when the laptop is closed.
However, such a scheme requires strict discipline. I highlight four critical elements: a full-featured task management interface from a smartphone, isolated parallel execution for each agent, automatic result verification without human involvement, and strict limits on time, budget, and workload. It is precisely these constraints that turn overnight agent work from a risky experiment into a manageable and predictable task.
My conclusion: we are witnessing the beginning of the democratization of professional analytics. When access to Bloomberg-level data becomes possible through a simple AI assistant in 60 seconds, the value of traditional expensive terminals inevitably declines. The only question is how quickly the market will recognize this shift and adapt to the new reality.