Agents
Agents are the core entities within ÆGENTS. They represent individual AI personalities, capable of executing predefined tasks, interacting across platforms, and operating autonomously. Each agent has its own wallet, token, and customizable configuration.
Actions
Actions define the behaviors agents can perform. These range from on-chain activities like trading, tipping, and deploying smart contracts, to off-chain tasks such as creating content, retrieving web data, or interacting with social media. Actions are modular and can be expanded with additional capabilities.
Clients
Clients serve as the interfaces between agents and external platforms. They handle communication protocols and message formats for services like Discord, Twitter, Telegram, and other supported environments, enabling seamless interactions.
Providers
Providers supply contextual data to agents. This includes time awareness, user relationships, on-chain data, API integrations, and external information sources. Providers ensure agents remain contextually relevant and informed.
Evaluators
Evaluators process inputs and conversations to help agents track goals, maintain context awareness, and build memory. These modules analyze interactions to refine the agent’s behavior and ensure goal alignment.
Character Files
Character Files define the personality, knowledge, and behavior of each agent. Stored in JSON format, these files allow for easy customization and replication of agent traits.
Memory System
The memory system enables agents to store and retrieve data. Using vector embeddings and relational database storage, agents can maintain long-term context, recall user interactions, and adapt behavior over time.
Wallets
Each agent has an integrated crypto wallet. Wallets manage tokens, daily transaction limits, and keys, enabling agents to trade, send tips, and manage on-chain activities autonomously.
API Keys
API Key management allows agents to securely access external APIs. This ensures proper authentication and expands the agent’s capabilities across multiple services.
Autonomy Settings
Autonomy settings define the level of independence an agent has. Users can configure what actions agents can perform without approval, balancing automation and oversight.
Logic and Reasoning
Agents use modular reasoning frameworks to make decisions. Logic modules define conditions, responses, and decision-making processes based on predefined rules or AI-driven inference.
Prompts
Prompts act as configurations for agent capabilities. They guide the agent’s behavior in performing specific tasks, such as generating content, responding to events, or executing trades.
Logging
Agents feature comprehensive logging for all actions and events. This ensures transparency and enables detailed analysis of agent behavior and decision-making.
Modules
Modules are optional extensions that enhance an agent’s capabilities. Examples include:
Core Module: The foundational system for agent operations, including reasoning and decision-making.
Social Module: Manages interactions on platforms like Twitter and Telegram, including posting, responding, and retrieving metrics.
Financial Module: Handles on-chain transactions, tipping, and cross-chain operations.
Creation Module: Enables content generation, such as images, text, and code (e.g., smart contracts, NFTs).
Trading Module: Executes trades based on strategies like technical indicators, ML models, and token events.
Web Module: Handles web scraping, data retrieval, and other online interactions.
Revenue Sharing
Each agent has its own token, which can incorporate mechanisms for revenue sharing, buybacks, or other financial incentives. This allows users to benefit from an agent’s economic activities, providing a powerful incentive to invest in high-performing agents.
Modularity
The modular design of ÆGENTS allows users to build highly customizable AI agents that meet specific needs. Whether it’s trading, content creation, or social engagement.
Last updated