Gemini Launches Agentic Trading: Claude and ChatGPT Can Now Run Your Crypto Account

Gemini opened its trading APIs to AI agents Monday via Anthropic's MCP open standard, letting Claude and ChatGPT autonomously place trades, monitor markets, and manage risk. Gemini calls it the first agentic trading tool through a regulated US exchange.

Gemini agentic trading cover — Claude ChatGPT MCP integration for crypto
Illustration: BlockAI News · Source: Gemini announcement / The Block, April 27 2026

Crypto exchange Gemini on April 27 launched Agentic Trading, a new feature that lets users connect AI models — including Anthropic's Claude and OpenAI's ChatGPT — directly to their trading accounts via the open Model Context Protocol (MCP). Gemini is calling it the first agentic trading tool available through a regulated US exchange.

The Launch

Connected agents can autonomously monitor markets, place orders, and execute risk-management strategies against predefined rules. Gemini exposes three core "skills" through MCP: Get Market Data (live price and order-book data), Find the Spread (bid-ask checks), and Retrieve Candles (historical price series). Both Claude and ChatGPT are supported out of the gate, and Gemini says more skills will land in coming weeks.

How It Works

MCP, released by Anthropic in late 2024 and since adopted by OpenAI, Google DeepMind, Microsoft, and others, gives AI agents a standardized way to call external tools. The standard means a Claude- or ChatGPT-connected agent doesn't need a custom Gemini integration to trade — it just needs MCP support. Coinbase has rolled out a parallel Payments MCP tool focused on stablecoin transfers and onramps; Gemini's product goes further by giving agents real trade-execution authority.

Gemini rolls out Agentic Trading allowing AI bots to directly manage crypto exchange trading accounts
The Block on the launch, MCP standard, and how it compares to Coinbase's Payments MCP.

Why It's Different

The thing to watch is liability when an agent goes wrong. Gemini is a regulated entity; if a Claude agent slips on a malicious prompt and dumps a user's portfolio, the legal questions are unsettled — the user authorized the agent, but the agent acted on instructions buried in a third-party page or memory. Expect this to become the test case for retail agentic finance, and expect insurers and regulators to weigh in fast.

Daily Web3 + AI signal, straight to your inbox — subscribe →

Keep Reading

Ineffable Intelligence Raises $1.1B Seed at $5.1B as DeepMind's David Silver Bets on No-Human-Data AI

Ineffable Intelligence Raises $1.1B Seed at $5.1B as DeepMind's David Silver Bets on No-Human-Data AI

British AI lab Ineffable Intelligence, founded only months ago by former DeepMind reinforcement-learning chief David Silver, raised a $1.1 billion seed round at a $5.1 billion valuation on April 27. The round — Europe's largest seed ever — was co-led by Sequoia and Lightspeed, with Nvidia, DST Global, Index Ventures, Google, and the UK's Sovereign AI Fund participating.

Inside the Round

Sequoia and Lightspeed taking co-lead at this scale signals US capital is increasingly comfortable backing European frontier labs as long as the founder pedigree

Read full story →

Stay Ahead of the Market

Daily AI & crypto briefings — straight to your inbox, your phone, and your timeline.