Binance launches Agent OS to enable AI-driven trading with user-controlled limits

Binance, the world’s largest crypto exchange with more than 300 million registered users, has launched Agent OS, a platform that enables artificial intelligence agents to analyse markets and execute trades on behalf of users. The system connects AI applications to the exchange’s financial infrastructure, including its APIs, Wallet Agentic Hub, and x402 transaction verification services. By supporting the Model Context Protocol, Agent OS allows developers to integrate tools such as OpenAI’s ChatGPT and Codex, Anthropic’s Claude Code, and Cursor, granting agents access to market data and trading capabilities.

The platform marks a shift in the AI industry from conversational chatbots to autonomous agents capable of executing financial actions. Jeff Li, vice president of product at Binance, stated that the company places granular access control in users’ hands to protect funds. Users can configure dedicated sub-accounts for agents, specifying permissions for activities such as spot or futures trading. Withdrawals from these sub-accounts are blocked by default, creating a sandbox environment that limits an agent’s ability to move funds out of the designated space.

Binance does not impose separate caps on trading volume or losses within sub-accounts; the limit is effectively determined by the funds a user transfers into the account. Users can choose whether an agent must seek approval for every order or can execute trades autonomously once permissions are set. Li noted that while Binance can monitor resulting trading activity, it cannot see the reasoning behind an agent’s decisions, as this occurs outside its systems on the user’s computer or within their chosen AI application. This limitation means the exchange has restricted visibility into whether decisions were influenced by faulty information or manipulation.

To manage risk, Binance has set daily limits for Agentic Wallet transactions, which differ from exchange trading caps. Regular swaps are capped at $50,000 per day, DeFi transactions have a default limit of $100,000, and x402 payments are restricted to $20 daily. The company stated that its existing security, risk-control, and anti-money-laundering policies for sub-account APIs apply to Agent OS at launch. Li described the platform as a first step toward enabling developers to build AI-powered applications that can act across crypto and traditional markets.

The launch follows a broader trend among rival exchanges to adopt similar technologies. Kraken launched an open-source command-line tool with Model Context Protocol support in March, allowing agents to execute spot and futures trades. Coinbase introduced Coinbase for Agents in June, connecting AI agents directly to user accounts for trading and payments. OKX also enabled agentic trading earlier in the year by integrating an open-source MCP toolkit, reflecting a competitive push to integrate autonomous AI into financial infrastructure.