Keypo is a new primitive that combines authentication, permissioning and data encryption into a single composable object. Humans and machines use Keypo to securely share and transact with digital secrets (information that should be secured like login credentials, code snippets, API keys, premium content, etc).
The primitive uses distributed ledger technology and smart contracts to create a universal key management system that is secure, open, and standard. It supports a variety of use cases, including:
- AI agents can access APIs and pay for usage with no human in the loop.
- Flexible, secure crypto wallet sharing across multiple human and machine participants.
- Machine connectivity to IOT and other hardware systems.
- Flexible agent - agent coordination.
AI agents today depend on centralized platforms for access to third-party services, making credential management inefficient and insecure. Keypo eliminates this friction by allowing agents to autonomously acquire and use tokenized access rights—such as API keys and passwords—without ever handling raw credentials. Through secure hardware and crypto wallets, agents can transact trustlessly, enabling seamless machine-to-machine interactions.
Keypo Architecture
https://www.figma.com/board/YVUAl6oBYzbtw7Rh5FAuww/Protocol-Block-Diagram-EXTERNAL?node-id=0-1&t=Qu0sfZDGfn3Zj4NL-1
Key Properties
- Storage Provider Agnostic
- Users can store their encrypted secrets wherever they want (IPFS, S3, Dropbox, Google Drive, private local servers, etc.).
- Why this is useful:
- Users can move encrypted secrets between storage providers without reestablishing access conditions. Encrypted secrets can be stored across multiple storage providers, ensuring uninterrupted access even if a provider goes offline.
- Decentralized storage options mitigate platform risk.
- Public, Decentralized Permissioning
- Access rules are stored on-chain, eliminating reliance on centralized permissioning.
- Why this is useful:
- Multiple applications share the same data layer, meaning file permissions are portable across applications.
- Using blockchains for permissioning vs centralized services like AWS significantly reduces platform risk.
- Machines can verify and request access without intermediaries, enabling autonomous transactions.
- Distributed Encryption and Decryption Using Multiparty Computation (MPC)
- The protocol relies on distributed MPC networks for secure, decentralized encryption and decryption, removing the need for centralized key management.
- Why this is useful:
- Secrets remain secure—no single point of failure or key exposure.
- Higher portability across applications vs. centralized alternatives and lower vendor/platform risk.
- Machines can decrypt and utilize secrets autonomously without ever possessing raw credentials, ensuring security in machine-to-machine interactions.
Example Use Cases
- Selling access to an API
- Upload an API key using the Keypo client.
- Set a price and define permission rules (e.g., discounted access for power users or time-limited trials for new users).
- Users can purchase access on-chain and make API requests without ever handling raw credentials.
- AI assistant accessing a user's cloud storage securely.
- Upload cloud storage credentials using the Keypo client.
- Grant your AI assistant permission to use credentials, identified by their wallet address or other crypto-native methods.
- When the assistant needs to access cloud storage, it retrieves uses the credentials without ever seeing the raw keys.
- Releasing content directly to fans
- Upload content using the Keypo client.
- Set the parameters for distributing the album. You can set any price (including $0) and put restrictions on who can buy it. For example, you may want to let purchasers of previous releases buy it at a discount.
- Anyone can log in to a compatible client and view all files available for sale.