MoonPay has published an open-source wallet standard aimed at allowing AI agents to manage funds and carry out transactions across multiple blockchain networks. The announcement, made on Monday, addresses a significant gap in how autonomous software systems interact with cryptocurrency infrastructure. Currently, many setups require each system to manage its own keys and balances separately, creating fragmented and inefficient arrangements. The new standard replaces this approach by giving AI agents access to a unified pool of funds across tools and blockchains.
The specification stores private keys in an encrypted local vault and handles transaction signing within an isolated process, ensuring that keys remain outside the AI agent’s active runtime environment. This design is intended to reduce security risks associated with autonomous software handling sensitive cryptographic material. The system also incorporates policy controls that allow users to define spending limits and other restrictions before any transaction receives approval. These safeguards are meant to give developers and end users greater oversight of how agents spend funds.
MoonPay noted that while recent industry efforts have concentrated on transaction rails for machine-driven payments, they have largely overlooked how wallets and private keys are actually managed. Researchers at MIT Sloan have described AI agents as capable of using standard building blocks, such as APIs, to communicate with other agents and humans, send and receive money, and interact with the internet. The new standard is designed to support exactly this kind of interoperability. It is modular in structure, with components covering storage, signing, policy controls and chain support.
The specification is available through developer platforms including GitHub, npm and PyPI. More than a dozen companies contributed to its development, among them PayPal, OKX and Circle, along with several blockchain foundations and infrastructure providers. MoonPay, founded in 2019, provides financial technology infrastructure enabling businesses and consumers to move funds between fiat currencies and digital assets, offering services such as on- and off-ramps, trading and crypto payments across global markets.
In a separate Monday announcement, BitGo, a digital asset custody and infrastructure company, said it had launched a Model Context Protocol server allowing AI-driven tools to access its developer platform using natural language. The integration enables agents to navigate wallet functions, transaction flows and staking systems. It also connects BitGo’s infrastructure to AI-native development environments, allowing tools such as ChatGPT and code editors to retrieve documentation, API references and product information directly within their workflows.
These developments reflect a wider industry push to embed crypto services into AI systems, as companies explore ways for software to engage with financial infrastructure without depending on conventional user interfaces. Other initiatives in this space include Coinbase‘s x402 protocol, which facilitates stablecoin transfers over HTTP for APIs, applications and AI agents. Tools introduced last week by Visa and Stripe-backed Tempo similarly allow AI systems to initiate and execute payments programmatically. Taken together, these efforts signal growing momentum behind the concept of AI agents functioning as independent economic participants within crypto ecosystems.
Originally reported by CoinTelegraph.
