Struct OpenAIRequest 

pub struct OpenAIRequest {
    pub model: String,
    pub messages: Vec<OpenAIMessage>,
    pub stream: bool,
    pub max_tokens: Option<u32>,
    pub temperature: Option<f32>,
    pub tools: Option<Vec<Value>>,
}

Complete request payload for OpenAI chat completions API.

This struct is serialized and sent as the request body when making API calls to OpenAI-compatible endpoints. It includes the model, conversation history, and configuration parameters.

Streaming

The SDK always uses streaming mode (stream: true) to enable real-time response processing and better user experience.

Optional Fields

Fields marked with skip_serializing_if are omitted from the JSON payload when None, allowing the API provider to use its defaults.

Example

use open_agent_sdk::types::{OpenAIRequest, OpenAIMessage};

let request = OpenAIRequest {
    model: "gpt-4".to_string(),
    messages: vec![
        OpenAIMessage {
            role: "user".to_string(),
            content: "Hello!".to_string(),
            tool_calls: None,
            tool_call_id: None,
        }
    ],
    stream: true,
    max_tokens: Some(1000),
    temperature: Some(0.7),
    tools: None,
};

Fields

model: String

Model identifier (e.g., “gpt-4”, “qwen2.5-32b-instruct”).

messages: Vec<OpenAIMessage>

Conversation history as a sequence of messages.

Includes system prompt, user messages, assistant responses, and tool results. Order matters - messages are processed sequentially.

stream: bool

Whether to stream the response.

The SDK always sets this to true for better user experience. Streaming allows incremental processing of responses rather than waiting for the entire completion.

max_tokens: Option<u32>

Maximum tokens to generate (optional).

None uses the provider’s default. Some providers require this to be set explicitly.

temperature: Option<f32>

Sampling temperature (optional).

None uses the provider’s default. Controls randomness in generation.

tools: Option<Vec<Value>>

Tools/functions available to the model (optional).

When present, enables function calling. Each tool is described with a JSON schema defining its parameters. None means no tools are available.

Trait Implementations

impl Clone for OpenAIRequest

fn clone(&self) -> OpenAIRequest

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for OpenAIRequest

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Serialize for OpenAIRequest

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.