1. What is HAI (Hiro the hAIpe)

1.1 Background and Origin of HAI

1.1.1 Definition and Core Concepts of HAI

HAI (Hiro the hAIpe) is a blockchain-based AI agent project that focuses on the HyperLiquid ecosystem. Its core concept is to combine AI technology with blockchain in a decentralized manner to provide users with transparent and efficient AI services. The goal of HAI is to become a leader in the field of AI agents and promote the democratized application of AI technology.

1.1.2 Developer of HAI: Virtuals Team

HAI is developed by the Virtuals team, which has extensive experience in the blockchain and AI fields. The Virtuals team is committed to promoting the development of decentralized applications (DApps) through technological innovation. HAI is an important attempt by the Virtuals team in the field of AI agents, aiming to solve the trust and transparency issues in traditional AI services through blockchain technology.

1.1.3 The inspiration and naming of HAI

The naming inspiration for HAI comes from the combination of the words ‘Hiro’ and ‘hAIpe’. ‘Hiro’ represents the protagonist of the project, the AI agent; ‘hAIpe’ combines ‘AI’ and ‘hope’, symbolizing hope and innovation brought by AI technology. The design concept of HAI is deeply influenced by internet culture and community-driven models, aiming to attract users’ participation through interesting and easy-to-understand methods.

2. HAI’s Technical Architecture

Now introducing the agent framework developed by hAIpe and SIFU. The specific SIFU Agent framework is as follows:

The goal of the SIFU Agent framework is to build an entity with the ability to perform self-executing tasks, receive external information, process internal logic, and manage states. All functionalities are implemented in the Node.js environment using TypeScript syntax. This framework is designed to be modular, flexible, and extensible, allowing developers to customize and optimize different functional modules as needed.

2.1 Core Module

State: Stores information for the agent, including short-term context, long-term memory, configuration parameters, etc. Tasks and thoughts can read or update the state to provide the key information required for the agent.

Task: represents the actions that an agent needs to actively perform, which may include certain execution processes and potential outputs. Tasks may require cognitive processing or may read/update data from the state.

Thinking: A set of universally callable interfaces that represent the proxy’s reasoning capability. Simply provide the input context, the thinking module can return the decision or result. Thinking can encapsulate complex logic, such as interaction with large language models (LLMs) or integration with multiple rule engines, and can access states for reasoning.

Sensing: Responsible for monitoring the event pool and responding to external events, the core module. It triggers tasks, updates status, or creates new events based on external data.

2.2 Auxiliary Components

Event Pool: By subscribing to events in the event pool, Sensing listens for specific events and responds accordingly.

Event Pump: The event pump injects events into the event pool based on specific conditions (such as scheduled tasks, Webhook triggers, etc.).

Reflex: The reflex mechanism allows the agent to quickly respond and utilize its state and cognitive abilities based on input-driven interaction, without the need for task invocation or perception. This mechanism is inspired by the nervous system and is used for rapid, reflexive responses.

2.3 State Data Storage

The State module is responsible for storing various types of data, including:

• Internal Memory: The agent’s long-term memory and important data.
• Parameters: configuration options and intermediate results used in tasks and thinking.
• Knowledge Base: A structured information repository used for inference and querying.

Multi-level differentiation

State is divided into multiple levels:

• Short-term context: temporary, immediate status information, usually used to deal with current tasks or thinking needs.
• Long-term memory: data or experiences accumulated over time, usually affecting reasoning or decision-making.
• Configuration parameters: the settings or tuning parameters of the framework to ensure the flexibility of tasks and thinking.

2.4 Modular Management

Each module can maintain its own state and read or write to the main state. This modular approach combines state management with the design of each module, improving scalability and flexibility.

For long-running agents or those that need to preserve historical information after a restart, the state can be periodically saved to a database or file, or stored only in memory.

The SIFU Agent framework is designed with modularity and flexibility in mind, supporting seamless integration of core modules such as state management, task execution, reasoning, and perception. It adopts an event-driven architecture for efficient internal and external communication through event pools and event pumps. The reflection mechanism enables rapid response and reduces the need for complex logic execution.

By clearly separating concerns and supporting scalability, SIFU Agent provides a powerful foundation for building automated systems that can handle information, perform tasks, and make decisions in dynamic environments.

3. Basic Information of HAI Token

3.1 Basic Information of HAI Token (2025-2-5)

1. Market Cap: $86.44K
2. Circulation: 10.00 billion HAI
3. Fully diluted market cap: $86.44 million
4. Total Supply: 10.00 billion HAI
5. Public Chain: BASE Chain
6. Contract Address: 0xeb70ce81f69ec4b7e425e438e9f5fb78d95b50ec

Market Performance of 3.2 Tokens

Currently, the HAI token has been listed on the Gate.io innovation zone.Click to trade!

Risk Warning: Cryptocurrency projects may have high volatility and high risk. Please trade cautiously and be aware of the risks!

Conclusion

HAI (Hiro the hAIpe) is an innovative project that combines AI agents with blockchain technology and has broad application prospects. Its unique technical architecture, economic model, and community-driven approach provide users with transparent and efficient AI services. Whether it can occupy an important position in the fiercely competitive cryptocurrency market in the future depends on its engineering capabilities, and whether it can ensure the implementation and user experience of its products.