Struct AsyncLlmAgent 

pub struct AsyncLlmAgent { /* private fields */ }

An async agent powered by an LLM.

This agent uses an LLM broker to generate responses. It can be configured with a system prompt (behaviour), optional tools, and a response model for structured output.

Examples

use mojentic::agents::AsyncLlmAgent;
use mojentic::llm::LlmBroker;

let broker = Arc::new(LlmBroker::new("model-name", gateway, None));
let agent = AsyncLlmAgent::new(
    broker,
    "You are a helpful assistant.",
    None, // tools
);

Implementations

impl AsyncLlmAgent

pub fn new( broker: Arc<LlmBroker>, behaviour: impl Into<String>, tools: Option<Vec<Box<dyn LlmTool>>>, ) -> Self

Create a new AsyncLlmAgent.

Arguments

• broker - The LLM broker to use for generating responses
• behaviour - System prompt defining the agent’s personality and behavior
• tools - Optional tools available to the LLM

pub fn add_tool(&mut self, tool: Box<dyn LlmTool>)

Add a tool to the agent.

Arguments

• tool - The tool to add

pub async fn generate_response( &self, content: &str, correlation_id: Option<String>, ) -> Result<String>

Generate a text response using the LLM.

Arguments

• content - The user message content
• correlation_id - Optional correlation ID for tracing

Returns

The generated text response

pub async fn generate_object<T>( &self, content: &str, correlation_id: Option<String>, ) -> Result<T>

where T: for<'de> Deserialize<'de> + Serialize + JsonSchema + Send,

Generate a structured object response using the LLM.

Arguments

• content - The user message content
• correlation_id - Optional correlation ID for tracing

Returns

The generated structured object

Trait Implementations

impl BaseAsyncAgent for AsyncLlmAgent

fn receive_event_async<'life0, 'async_trait>( &'life0 self, _event: Box<dyn Event>, ) -> Pin<Box<dyn Future<Output = Result<Vec<Box<dyn Event>>>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait,

Process an event asynchronously and return resulting events.