Trait Provider 

pub trait Provider:
    Send
    + Sync
    + Debug
    + 'static {
    // Required methods
    fn info(&self) -> Arc<ProviderInfo>;
    fn chat_completion<'life0, 'async_trait>(
        &'life0 self,
        req: ChatCompletionRequest,
    ) -> Pin<Box<dyn Future<Output = Result<ChatCompletionResponse, AiError>> + Send + 'async_trait>>
       where Self: 'async_trait,
             'life0: 'async_trait;
    fn stream_chat_completion<'life0, 'async_trait>(
        &'life0 self,
        req: ChatCompletionRequest,
    ) -> Pin<Box<dyn Future<Output = Result<Box<ChatCompletionStream>, AiError>> + Send + 'async_trait>>
       where Self: 'async_trait,
             'life0: 'async_trait;
}

Core provider trait for AI services.

This trait defines the simplified interface that all AI providers must implement. Providers only need to implement the basic chat completion API, and higher-level abstractions (generate_text, generate_object) are handled by the Runtime layer.

Required Methods

fn info(&self) -> Arc<ProviderInfo>

Get provider information

fn chat_completion<'life0, 'async_trait>( &'life0 self, req: ChatCompletionRequest, ) -> Pin<Box<dyn Future<Output = Result<ChatCompletionResponse, AiError>> + Send + 'async_trait>>

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

Chat completion (non-streaming)

This is the core method that all providers must implement. It handles a chat completion request and returns a response.

fn stream_chat_completion<'life0, 'async_trait>( &'life0 self, req: ChatCompletionRequest, ) -> Pin<Box<dyn Future<Output = Result<Box<ChatCompletionStream>, AiError>> + Send + 'async_trait>>

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

Stream chat completion

Returns a stream of chat completion chunks for streaming responses.

Implementors