Struct LlmClient 

pub struct LlmClient { /* private fields */ }

Implementations

impl LlmClient

pub fn from_config( config: &RobitConfig, profile_name: Option<&str>, ) -> Result<Self, LlmError>

Create a new LlmClient from loaded configuration.

profile_name: which profile to use. If None, uses the default profile.

pub async fn chat_stream( &self, messages: Vec<ChatCompletionRequestMessage>, tools: Option<Vec<ChatCompletionTools>>, ) -> Result<ChatCompletionResponseStream, LlmError>

Streaming chat completion. Returns an async stream of response chunks.

pub async fn chat( &self, messages: Vec<ChatCompletionRequestMessage>, tools: Option<Vec<ChatCompletionTools>>, ) -> Result<CreateChatCompletionResponse, LlmError>

Non-streaming chat completion. Returns the full response.

pub fn model(&self) -> &str

Get the current model ID (e.g. “deepseek-chat”).

pub fn profile(&self) -> &str

Get the profile name (e.g. “default”).

pub fn resolved(&self) -> &ResolvedModel

Get the resolved model info.

pub fn supports_images(&self) -> bool

Whether the current model supports image inputs.

pub fn supports_tools(&self) -> bool

Whether the current model supports tool calling.