T-Investments MCP server — AI agents for market and portfolio analysis
· Source: original
T-Investments MCP server: a brokerage API for AI agents without Python scripts
🤖
T-Investments has launched an MCP server through which AI agents work with the brokerage API directly. MCP stands for Model Context Protocol, a standard layer between LLMs and a broker's API. The server is public.
What the agent can do straight from the model: find suitable securities, analyze the market and the client's portfolio, get portfolio status, view the order book, place and cancel orders. Without manual Python scripts.
The server's client can be an AI agent like Hermes Agent, an automation platform like n8n, a desktop assistant, or an IDE. Previously, even a simple trading algorithm dragged along its own script: Python, the broker's API, cron, and error handling. For a retail investor, the barrier to entry into professional trading systems remained high.
The author of the breakdown is Alexey Kabanov, a developer on the public API team at T-Investments. In the article, he shows how to build a Telegram assistant and a trading robot on the T-Investments MCP without manually writing Python scripts — the breakdown on Habr.
With a minimal edit: only the hook was replaced. The previous one ("T-Investments rolled out a public MCP server for AI agents") repeated the already-used pattern of "company + colloquial action verb + product" — the same construction as in the posts "Google rolled out Gemini 3.8 Live with Live Avatar" and "Tiny-LLM from PrismML rolled in..." (that is, one colloquial flavor and one pattern with the company at the start), and it began the same way as the nearest post about Muse — with a company-action. The new hook is built on the "product + result" scheme, does not duplicate any of the last three schemes (product + verb in Gemini 3.8 Live, the numeral roundup in the post about 10 agents, the prepositional opening in the post about Muse), and does not repeat the final resolution. The tone of the post is neutral-analytical; among the last three there is no third consecutive enthusiastic or skeptical post, so the tone was not changed. The 🤖 emoji, paragraph structure, closing link, voice, and facts are unchanged
🤖 Interested in AI agents and automation?
Prompts for building AI agents and automations — read up on the topic:
🔗 The entire prompt library · The "AI Agents" category
A ready-made product on the topic: AI Agent Skills Pack — grab it and apply it right away.