Struct LlmClient 

pub struct LlmClient { /* private fields */ }

Unified LLM client — routes to Claude API, Grok API, or Ollama based on model name. Supports both blocking and streaming generation.

Implementations

impl LlmClient

pub fn new(model: &str) -> Self

pub fn with_limits(model: &str, context_size: u32, max_predict: u32) -> Self

pub async fn generate( &self, role: &str, system: &str, user_prompt: &str, ) -> Result<String>

Generate text (blocking — waits for full response).

pub async fn generate_with_stats( &self, role: &str, system: &str, user_prompt: &str, ) -> Result<(String, LlmCallStats)>

Generate text and return stats for reports.

pub async fn generate_live_with_stats( &self, role: &str, system: &str, user_prompt: &str, ) -> Result<(String, LlmCallStats)>

Generate with live streaming and return stats for reports.

pub async fn generate_live( &self, role: &str, system: &str, user_prompt: &str, ) -> Result<String>

Generate with live token-by-token output to stdout. Shows the model’s thinking process in real-time.

pub async fn generate_streaming( &self, role: &str, system: &str, user_prompt: &str, tx: Sender<StreamEvent>, ) -> Result<String>

Generate text with streaming — tokens sent via channel as they arrive. Returns the full accumulated text when done.

pub async fn chat_with_tools( &self, messages: &[ChatMessage], tools: &[OllamaTool], ) -> Result<ChatToolResponse>

Chat with native tool calling. Returns assistant content + tool calls.

Trait Implementations

impl Clone for LlmClient

fn clone(&self) -> LlmClient

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.