Magic Labs Debuts Verifiable Crypto Agent Platform Magic Newton

Leading wallet infrastructure provider Magic Labs announces the Genesis launch of Magic Newton, an AI agent platform designed to simplify crypto interactions for users and enable safe, autonomous finance. As the first verifiable crypto copilot, Newton reduces the complexity of managing digital assets by allowing users to delegate and automate tasks to intelligent onchain agents, while maintaining total control. Magic Newton is powered by the Newton Protocol that serves as trust infrastructure, ensuring every action is safe and transparent. Newton’s initial launch on Base will allow users to interact on the platform for the first time and try out native agents developed by Magic Labs.

Magic Labs, backed by PayPal Ventures and known for its secure onboarding of over 50 million embedded wallets via its many partners— such as Polymarket, Wallet Connect, and Naver, among others— has created Magic Newton to further improve crypto UX via trusted AI automation. Currently, users must either surrender private keys or create separate wallets for AI agents, options that compromise security and usability. Newton reimagines this relationship by making security the foundation for application abstraction, replacing the need to navigate complex interfaces with a simple command center. Users simply tell AI agents what they want to accomplish like “stake my tokens for the best yield”, and the agents help securely execute these tasks across multiple blockchains, with cryptographic proofs ensuring they’re following the user’s exact instructions.

The Genesis launch marks the first step in Magic Newton’s roadmap to transform how AI agents build trust with everyday users. It introduces what Magic Labs calls its “black triangle moment”, a reference to early PlayStation development where rendering a simple black triangle proved the entire graphics pipeline was functioning. This initial demonstration will showcase Newton’s integrated automation stack and streamlined user interface through easy yet powerful use cases, including automated dollar-cost averaging. It will also serve as proof that the entire verifiable automation stack is fully integrated and operational.

“What we’ve learned from onboarding 50 million wallets is that removing friction is only valuable if you maintain trust,” said Sean Li, co-founder and CEO of Magic Labs. “The real challenge is what happens after access—billions flow through fragmented chains daily, yet even advanced users struggle with complexity. Newton creates the trust layer for application abstraction, where users simply express their financial goals and AI agents execute as instructed within verifiable boundaries.”

Magic Newton is built with the end user in mind and takes a security-first approach to AI automation, focusing on trust and verification rather than just capabilities. The Newton Protocol, spearheaded by the Magic Newton Foundation and developed with key contributions from Magic Labs, will drive the verifiable automation layer for the Magic Newton ecosystem. The Protocol’s architecture is being built to address both sides of the security equation through a comprehensive “prevent-then-verify” approach:

• Proactive Protection: Before execution, Newton utilizes Smart Accounts that issue temporary “permission slips” (session keys) to AI agents, similar to how you authorize Google to access specific data, allowing them to perform actions within a user’s predefined boundaries, like maximum spending limits. Coming soon, Newton will enable more expressive permissions powered by zero-knowledge proofs where users can create conditions based on offchain real-time data, like “trade only if crypto sentiment is positive”.

• Retroactive Verification: After execution, all agent activity is validated through Trusted Execution Environments (TEEs), which generate cryptographic proofs showing that the agent’s code ran exactly as intended without manipulation. These proofs will be verifiable on-chain, creating an immutable record of all agent actions.

• Reputation Through History: Newton will establish a trust framework where agents earn reputation through proven, consistent, correct behavior over time. This system economically incentivizes reliability and penalizes misbehavior, creating a self-reinforcing ecosystem where the most trustworthy agents rise to the top through demonstrable performance.

Magic Newton’s layered security model makes it fundamentally different from other AI-powered financial tools. By combining preventative measures, cryptographic verification, and reputation management, Newton ensures that agents remain trustworthy even when executing complex financial actions across multiple blockchains.

“We share Magic’s vision for the future of agentic commerce: creating verifiable ways to leverage AI agents to handle complex processes so that consumers don’t have to. The launch of Magic Newton marks an important step forward in this rapidly emerging space,” said Alan Du, Partner at PayPal Ventures. “As developers harness the potential of agentic AI, we believe that simplifying the user experience by enabling agent instructions in plain language while maintaining robust security and verifiability will be critical to bringing trust to agentic AI.”

Looking ahead, Magic Labs plans to expand Newton into a broader community-driven marketplace where ecosystem participants can build and monetize secure agent-based automation services, enabling more advanced asset management strategies protected by the same cryptographic security guarantees.