Skip to main content

Warden AI Agents

Warden Agents​

A Warden AI Agent is an AI-driven program that supports both offchain and onchain operations.

Thanks to their native integration with Warden Protocol, Warden Agents can perform onchain Actions, such as creating Spaces, generating keys, executing transactions, and managing Orders.

You can easily build Agents with the Warden Agent Kit.

AI integration​

As explained in the sections below, Warden Agents' integration with AI is based on the following key components (previously called AIBI):

Warden allows inference endpoints to create cryptographic proofs that verify that an AI model produced a specific output. These proofs are then verified on Wardens’ blockchain through a new consensus mechanism, ensuring AI outputs are not just delivered, but can be trusted.

Within Warden, AI is integrated on two layers:

  • Verification layer
    Warden is designed to bring AIs onchain safely. Our verification layer uses blockchain, cryptography, and consensus to verify AI results are correct.

  • Execution layer
    Easily integrate AI with your application, protocols, and smart contracts. Automate their execution across any chain.

Key components​

Orders​

Warden Agents support the creation and management of Orders.

An Order is a Solidity smart contract performing onchain Actions at any destination chain. When creating an Order, you can implement custom logic by using Warden modules, Keychains, and other features.

In particular, the x/async module allows you to create smart contracts utilizing outputs of AI models. Orders can send any transactions to any Ethereum-based and EVM L2 application.

Learn more: Build an onchain AI Agent.

x/async​

The x/async module is a Cosmos SDK module for running offchain heavyweight computations asynchronously and storing the results onchain. It uses the ABCI 2.0 framework and its vote extensions to implement Tasks, Plugins, and Prophets.

The x/async module decouples transaction processing from consensus and moves away from the traditional wait-and-process model. Developers can initiate various crosschain computations including AI inferences and other tasks as asynchronous intents that are resolved later. This approach prevents smart contracts from blocking operations.

Learn more: x/async.

SPEX​

Statistical Proof of Execution (SPEX) is a sampling-based verifiable computing protocol that ensures the integrity of computational tasks through probabilistic guarantees. This includes tasks with potentially non-deterministic outputs, such as those involving large language models (LLMs) or stochastic training pipelines.

SPEX is used as a verifiability layer for AI on Warden Protocol. At the moment, SPEX verifies AI-driven price predictions generated by x/async. However, you can extend the module with other computational tasks and use SPEX to protect them.

Learn more: SPEX.