Crypto Agents

Embracing Verifiable Autonomy For Human Empowerment

Autonomy in technology often raises concerns—visions of uncontrolled AI, systems operating beyond human oversight, and unpredictable futures. However, when thoughtfully designed, autonomy can strengthen systems, making them more tamper-resistant and efficient. Autonomous systems like crypto agents serve as powerful allies in building resilient infrastructure. They can operate transparently, respond predictably, and function independently without relying on a central controller—traits that make them resistant to tampering and ideal for verifiable systems.

Rather than surrendering control to machines, autonomous agents empower users by providing environments where trust is embedded in the code itself. These systems protect against manipulation and streamline complex processes, making interactions seamless. While skepticism towards autonomy is understandable, recognizing its potential to empower and protect is equally important. By building autonomous, tamper-resistant agents within frameworks like Lit Protocol, developers can create a more secure, user-centric digital world.

What Is a Crypto Agent?

Not all autonomous agent based systems are crypto agents, for example, Surtrac is a traffic control system that runs in Pittsburgh that has reduced waiting times at lights by 40% by making each sensor (e.g traffic light) in the network its own agent.

A crypto agent is an autonomous software entity operating within a blockchain or decentralized environment, utilizing cryptographic tools to perform actions, make decisions, and interact with on-chain or off-chain resources. Typically powered by AI, these agents are designed to be secure and verifiable. They leverage blockchains and networks like Lit Protocol to perform tasks such as:

• Signing transactions
• Executing smart contract functions
• Provable Interacting with Web2 platforms

—all without human intervention or dependence.

It's worth noting that an AI agent whose private key is controlled by it's developer might also be considered a crypto agent. However, similar to an upgradable smart contract, if someone can "pull the plug," it isn't truly autonomous.

Crypto Agent Examples

Below are examples of autonomous protocols. Some have established teams actively working on them, while others are fresh concepts.

Liquidity Orchestration for CEX on Chain

The team at Genius Terminal has developed a chain-abstracted bridge that uses Lit Protocol as a global solver. In this system, anyone can contribute liquidity without maintaining their own rebalancing infrastructure, minimizing the risks associated with centralized solvers. The solution supports arbitrary call data execution after solving for liquidity, enabling full cross-chain intent execution and abstraction.

Implementation

This is achieved by utilizing programmable keys from Lit Protocol for liquidity orchestration across chains and asset types. Lit Protocol acts as a global solver to coordinate liquidity efficiently. For detailed diagrams and explanations, refer to the Genius white paper and Lit's docs on writing logic for conditional signing.

Data Orchestration + Recommendation Protocol

Imagine a data storage and sharing framework that enables users and websites to share data with one another, enhancing personalization and recommendations and building up user owned data stores. User-owned identity protocols like Plurality Network provide the foundation for creating such a data sharing and recommendation protocol. Users who aggregate their identity can exchange data with websites—such as e-commerce sites or ad publishers—for improved recommendations. In return, websites contribute data back to users, enriching their datasets. The valuable aggregated data encourages both parties to participate in the protocol.

Implementation

To bring this concept to life, we start by connecting and sharing user data. User data can be fetched via OAuth, passwords, APIs, and extensions and then stored encrypted with Lit Protocol on decentralized storage solutions like Orbis. Users have selective disclosure controls to clean, process, and selectively share their data with websites. Here's a developer guide on using fetch() with Lit.

Next, a method—such as a browser tag— is requested to allows websites to share data with users by adding to their 'data backpack'. This mutual data sharing enhances the user experience and provides both parties with more data.

Benefits

This approach kickstarts user-owned data ecosystems, empowers users with control over their data, and enhances personalization across platforms, which websites generally desire for higher conversion rates.

Liquidity Loss Prevention Bot

Managing risks in crypto lending is crucial, especially concerning impermanent loss and rebalancing. Impermanent loss occurs when providing tokens to a liquidity pool, and the price ratio of the paired tokens changes from the time of deposit. For instance, if you deposit ETH/USDC and the price of ETH doubles, you would have been better off holding the tokens individually.

Rebalancing involves the automatic adjustment of token ratios to maintain the desired pool ratio, following mathematical formulas used in Automated Market Makers. Arbitrageurs help restore balance by buying undervalued tokens and selling overvalued ones.

Implementation

To prevent liquidity loss, strategies like range-bound liquidity provision and active management are employed. Providing liquidity within specific price ranges reduces exposure to extreme price movements. Active management involves monitoring price movements, adjusting positions proactively, and implementing stop-loss mechanisms.

By automating these strategies with Lit Actions, liquidity providers gain greater peace of mind. Lit Protocol can execute these actions autonomously, ensuring timely responses to market changes without the need for constant human oversight.

Incentivized Learning with Microtransactions

Educational platforms can leverage AI and blockchain technology to incentivize learning by rewarding users for correct answers. Users engage with AI-driven questions based on educational content, such as a biology textbook. Their answers are evaluated by an AI model like OpenAI's GPT, providing immediate feedback.

Implementation

When a user answers a question, the AI model returns a simple 'yes' or 'no' indicating correctness. These results are recorded on the blockchain for transparency. Correct answers trigger microtransaction payments to the user's wallet, offering a tangible incentive to learn.

Smart contracts handle the reward distribution based on the AI evaluations. Lit Protocol is utilized for connecting the LLM to the blockchain by managing and utilizing the LLM API privately and within defined rules. Here's an example of connecting OpenAI to chain.

Benefits

This system motivates users through financial incentives, making education more engaging and interactive. It leverages the scalability of blockchain and AI technologies to create a rewarding learning environment.
h/t @EmblemVault

Inference Data Marketplace

An emerging opportunity is to create a marketplace where data providers sell access to API-gated data based on microtransactions. This allows consumers (specifically, a user's AI agent) to access valuable data on demand, paying only for what they use, while providing data providers with a new revenue stream.

Implementation

Data providers offer APIs that deliver specific datasets or services. Access to these APIs is controlled through microtransactions facilitated by smart contracts on the blockchain. When a consumer requests data, a microtransaction is executed, granting access to the API for that specific request.

Lit Protocol plays a crucial role by handling the API key management and authentication required for accessing data. It ensures that only authorized users can access the data and that the API key is not exposed to anyone. Here's a Lit developer guide for managing an API key.

Benefits

This model enables data providers to monetize their data easily and securely. Data consumers benefit from flexible, pay-as-you-go access without the need for subscriptions or large upfront costs. The real time nature of this system is optimal for specific data needed for inference / RAG too. The use of onchain microtransactions and Lit Protocol ensures realtime payment and protection of sensitive information (e.g. API key)

What's Lit Protocol?

Lit Protocol is a programable, decentralized secret management network and open developer platform. By managing private keys and encryption keys, the Lit network enables any sensitive data to be used in a threadhold and blind compute environment, like signing with private keys.

The purpose of the network is to keep secrets safe and make it easy to use them with Lit Actions.

Lit Actions are serverless functions and Javascript logic that you can use to program with secrets and preform blind compute. Lit Actions can be stored on IPFS, gaining immutability or on chain, gaining update-ability.

The security design is defense in depth, which stacks two privacy technologies: MPC TSS + Secure Hardware. The main issue with MPC is collusion, which the TEE prevents. The main issue with a TEE is a single point of failure, which MPC prevents.

The TEE is also where you can run Lit Actions, including fetch(), which opens up the opportunity for agents that operate across web2 platforms with private APIs keys as well as web3 networks with crypto keys.

Our goal at Lit Protocol is to create the most powerful tool for developers who want true autonomy or users in control. Lit v0 was launched in Feb of 24 and updated to v0.1 in August. The network has processed >20 million requests for builders like Fox, Lens, and Genius. The launch of Lit v1 is set for early 2025.

With this system, you can connect any blockchain, network, and platform and make them composable with each other using programmable secrets (e.g. private key, API key) and logic run on the sensitive data that agent or users connect to.

If you're ready to dive in, get started here:

Have a question or want to chat about ideas?

It's a unique time right now and as machine intelligence continues to advance, building user-controlled, resilient, and autonomous systems with cutting-edge tools is an excellent way to serve people in our changing world. There's never been a better time to start!