MetaMask has unveiled a new self-custodial AI wallet that allows artificial intelligence agents to carry out cryptocurrency trades autonomously, marking the company's latest step into automated on-chain finance.
A New Kind of Wallet
The product, called Agent Wallet, is designed to let AI agents execute transactions directly on the blockchain without requiring users to approve each move manually. Crucially, the wallet retains its self-custodial nature, meaning users keep control of their private keys and underlying assets rather than handing them to a third party.
The setup is built around user-defined boundaries. Owners can specify limits on how much an agent can spend, what types of trades it can make, and other guardrails intended to prevent an autonomous system from acting outside its mandate.
AI can now trade on your behalf — but only within the lines you draw.
Betting on Autonomous Trading
The launch reflects a broader effort by MetaMask to position itself at the intersection of crypto and artificial intelligence. As AI agents become more capable of interpreting market data and reacting in real time, wallet providers are racing to offer infrastructure that supports machine-driven activity while preserving user safeguards.
Autonomous trading carries obvious appeal for users who want to capitalize on fast-moving markets without being glued to a screen. It also raises questions about risk management, since an agent operating around the clock could amplify both gains and losses.
Key features highlighted by the rollout include:
- Self-custodial control so users retain their keys
- User-defined spending and trading limits
- The ability for AI agents to execute on-chain trades independently
What It Signals
By embedding AI agents into a widely used wallet, MetaMask is signaling that automated crypto trading is moving from experimental territory toward mainstream tooling. The company's large user base gives it significant reach to test how everyday participants respond to delegating financial decisions to software.
Whether the approach gains traction will likely depend on how well the guardrails hold up in volatile conditions and how much trust users are willing to place in autonomous systems handling their money.