Gemini, a US-regulated cryptocurrency exchange, has announced the launch of Gemini Agentic Trading, which it describes as the first agentic trading tool available directly through a regulated US-based exchange.
The tool, accessible through Gemini's Developer Platform, enables traders and developers to connect AI agents, including Claude, ChatGPT, and other models, to Gemini's exchange via the Model Context Protocol (MCP), allowing those agents to place trades, monitor markets, and manage risk autonomously on behalf of users.
The integration covers the full Gemini API through MCP, meaning any action currently available via the API is accessible to a connected AI agent. Users can connect existing AI tools without rebuilding their setup, define a trading strategy through natural language, and let the agent execute it.
Trading Skills and user accessibility
Gemini has introduced Trading Skills, modular, pre-built functions that AI agents can call to execute specific actions. Skills available at launch include retrieving real-time prices and order book depth, querying bid-ask spreads for any trading pair, and accessing historical OHLCV candlestick data for pattern recognition and backtesting. Users can instruct their AI agent in natural language, for example, to issue a sell order when Bitcoin reaches a specified price or when the bid-ask spread meets a defined threshold, and the agent uses the appropriate Skill to act.
The platform is designed to serve both experienced quantitative traders building fully custom multi-leg strategies and traders newer to automation who want to start with pre-built Skills and scale up complexity over time. Gemini has indicated the Skills library will be expanded to support fully agentic trading across a broader range of functions and models.
Back in December 2025, Gemini also received a Designated Contract Market (DCM) licence from the U.S. Commodity Futures Trading Commission (CFTC). The approval authorised Gemini Titan to operate a regulated derivatives exchange and allows the company to begin providing prediction markets to customers in the US.
