Standards We Advocate
Open protocols that prioritize human welfare and enable trustworthy AI commerce.
The Three-Layer Trust Model
"Is this agent legitimate?"
"Did the user approve this?"
"Process the transaction"
These layers are complementary, not alternatives. Each solves a distinct problem.
Cryptographic mandates for user-authorized agent spending. Enables users to grant agents spending authority with specific limits and constraints.
Checkout sessions and payment delegation for AI agents. Allows agents to complete purchases on behalf of users via Stripe.
PKI-based agent identity verification. Ensures an AI agent is legitimate and verified, not a malicious actor.
Standard for connecting AI models to external tools and data sources. Enables Claude and other models to interact with APIs.
Communication protocol for AI agents to discover and interact with each other. Enables agent marketplaces and collaboration.
Our Position
We advocate for these standards because they prioritize user control and transparent authorization. Users should always know what their AI agents are doing on their behalf, and they should have granular control over permissions.
We don't believe in lock-in. These are open standards from major industry players, and we implement them faithfully. Our competitive advantage isn't in proprietary protocols - it's in the quality of our implementation and the trust infrastructure we build on top.
Want to learn more about our technical approach? Check out our research or reach out to discuss.