Money Ann introduced an AI proxy access system for developers and started linking large model tools directly to transactions, wallets and ability to pay. After being authorized by the user, the AI agent can read real-time patterns, search account balances and perform operations in multiple transactions of currency.
Agent OS integration multiple interfaces
The new system is called Binance Age OS. The goal is to make it easier for developers to access the encrypted trade scene for AI agents.
One of the core components is Binance MCP Server. The service is built on the Model Context Protocol, or MCP standard, to allow compatible AI applications to connect external tools in a uniform manner and to reduce the number of steps for users to manage their own local API keys. Currency is listed as having access to Claude, ChatGPT, Codex and VS Code.
Dealable but not open.
In accordance with the functional range disclosed by the currency, the AI agent, with access and authorization, may:
- Read real-time market data
- Query account balances
- Billing in cash, leverage, flash and futures products
However, François does not open the most sensitive funds to AI agents. The operation can only be conducted through a separate “Agentic sub-account”, which is isolated from the user's main account asset. At the same time, the access does not include a currency-designing authority, which means that AI agents cannot transfer funds to outside wallets.
It also reminds users that orders and transfers should be individually checked before they are confirmed. This means that the ultimate operational confirmation responsibility remains primarily with the user, rather than being left entirely to the agent for automatic execution.
The platform competes for the AI proxy entrance.
It also shows that the trading platform is accelerating the competition for the AI proxy entrance. As large-model tools move into payment, wallet and trading chains, the exchange wishes to establish standardized interfaces and rights controls before attracting developers to build applications around these capabilities.
Similar actions have taken place within the industry. Coinbase had previously introduced tools to allow AI agents to trade and pay and supported the start-up companies through the Base accelerator. Gemini also has her own proxy trade function online.
In addition to the transaction function, the other main line of competition is trust and authority control. The report mentions that MetaMask has launched the self-hosted AI wallet product and MoonPay is also setting up the proxy product. Ledger also works with MoonPay to try to cap spending on AI agents in hardware wallets.