Struct ResponsesClient 

pub struct ResponsesClient<T>
where
    T: Transport,
{
    pub model_name: String,
    pub api_key: String,
    pub scheme: String,
    pub host: String,
    pub base_path: String,
    pub transport: T,
    pub extra_tool_declarations: Vec<Value>,
    pub reasoning_effort: Option<String>,
    pub use_previous_response_id: bool,
    pub store: Option<bool>,
    pub meta: ProviderMeta,
    /* private fields */
}

Generic client over the OpenAI Responses API wire format (POST {base_url}/responses). Provider wrappers like [chat_openai] hold one of these as their inner state.

Fields

model_name: String

api_key: String

scheme: String

host: String

base_path: String

transport: T

extra_tool_declarations: Vec<Value>

Tool declarations injected by the wrapping provider on top of the ToolDeclarations view supplied by Chat. The wire layer drops these into tools[] opaquely — it does not know about OpenAINativeTool or any provider-specific trait.

reasoning_effort: Option<String>

use_previous_response_id: bool

store: Option<bool>

meta: ProviderMeta

Trait Implementations

impl<T> CompletionProvider for ResponsesClient<T>

where T: Transport,

fn complete<'life0, 'life1, 'life2, 'life3, 'life4, 'async_trait>( &'life0 mut self, messages: &'life1 mut Messages, tool_declarations: Option<&'life2 dyn ToolDeclarations>, options: Option<&'life3 ChatOptions>, structured_output: Option<&'life4 Schema>, ) -> Pin<Box<dyn Future<Output = Result<ChatResponse, ChatFailure>> + Send + 'async_trait>>

where 'life0: 'async_trait, 'life1: 'async_trait, 'life2: 'async_trait, 'life3: 'async_trait, 'life4: 'async_trait, ResponsesClient<T>: 'async_trait,

Run one completion step.

fn metadata(&self) -> Option<&ProviderMeta>