Skip to main content

Abstractions

Abstractions​

Naptha provides several key abstractions to help you build powerful distributed multi-agentsAI applications. Let's explore each component:

Agents Modules​

Agents are autonomous entities that can perform tasks using:

  • System Prompts - Define the agent's personality and core behaviors
  • Tools/Skills - Capabilities that agents can use to interact with external systems
  • Persona Module - Additional personality traits and characteristics loaded into the system prompt
  • Memory - Contextual storage enabling agents to maintain state and learn from interactions
#naptha-sdk AgentConfig
class AgentConfig(BaseModel):
    config_name: Optional[str] = "agent_config"
    llm_config: Optional[LLMConfig] = LLMConfig()
    persona_module: Optional[Union[Dict, BaseModel]] = None
    system_prompt: Optional[Union[Dict, BaseModel]] = None
Agent Configuration Details
  • The system_prompt defines core agent behavior
  • llm_config specifies which language model to use under the hood
  • persona_module is loaded as a dictionary and merged into the system prompt
  • Tools/Skills are currently part of agent modules (moving to separate modules soon)
  • Memory system will also become a separate module in future updates

Interact with the Agent Hub​

You can also use the CLI to explore available agents that you can run on a node:

naptha agents

Learn more about Agents →

Multi-Agent Orchestrators​

Coordinate multiple agents to solve complex tasks through defined interaction patterns and workflows. Perfect for:

  • Task decomposition
  • Parallel processing
  • Multi-agent collaboration
#naptha-sdk OrchestratorConfig
class OrchestratorConfig(BaseModel):
    config_name: Optional[str] = "orchestrator_config"
    max_rounds: Optional[int] = 5
info

Builders can then create their own custom OrchestratorConfigs. For example, here is what a custom orchestrator config looks like for MarketAgents:

        "orchestrator_config": {
            "config_name": "orchestrator_config_1",
            "num_agents": 10,
            "max_rounds": 2,
            "environment_order": [
                "group_chat",
                "auction"
            ],
            "environment_deployments": [
                {"name": "group_chat"},
                {"name": "auction"}
            ],
            "protocol": "acl_message"
        }

Interact with the Agent Orchestrator Hub​

You can also use the CLI to explore available agent orchestrators that you can run on a network of nodes:

naptha orchestrators

Environments:​

Environments in Naptha provide the essential infrastructure that powers multi-agent interactions. They serve as the shared operational space where agents can:

  • Exchange information seamlessly
  • Maintain persistent state across executions
  • Access shared resources and context
  • Coordinate complex workflows
info

Similar to OrchestratorConfig, builders can inherit from the base to create their own custom environment config. Here is an example of the current one used for a MarketAgents module:

{
    "config_name": "groupchat",
    "environment_type": "groupchat",
    "max_rounds": 5,
    "initial_topic": "Initial Market Discussion",
    "sub_rounds": 3,
    "group_size": 5
}

Interact with the Environment Hub​

You can also use the CLI to explore available environments that you can use with orchestrators:

naptha environments