Struct OpenAiProvider 

pub struct OpenAiProvider { /* private fields */ }

Client for the OpenAI Chat Completions API (POST {base_url}/chat/completions).

The same wire format is served by many other providers; use the presets (azure, bedrock, groq, mistral, deepseek, xai, openrouter, together, fireworks, perplexity, ollama, lm_studio) or with_base_url for anything else (vLLM, LiteLLM, …).

Implementations

impl OpenAiProvider

pub fn new(api_key: impl Into<String>) -> Self

pub fn azure(resource: &str, api_key: impl Into<String>) -> Self

Azure OpenAI v1 API: https://{resource}.openai.azure.com/openai/v1, key in the api-key header. Pass the deployment name as the model.

pub fn bedrock(region: &str, api_key: impl Into<String>) -> Self

Amazon Bedrock, OpenAI-compatible endpoint with a Bedrock API key: https://bedrock-runtime.{region}.amazonaws.com/openai/v1. Use Bedrock model IDs.

pub fn groq(api_key: impl Into<String>) -> Self

Groq: https://api.groq.com/openai/v1.

pub fn mistral(api_key: impl Into<String>) -> Self

Mistral AI: https://api.mistral.ai/v1.

pub fn deepseek(api_key: impl Into<String>) -> Self

DeepSeek: https://api.deepseek.com/v1.

pub fn xai(api_key: impl Into<String>) -> Self

xAI (Grok): https://api.x.ai/v1.

pub fn openrouter(api_key: impl Into<String>) -> Self

OpenRouter: https://openrouter.ai/api/v1.

pub fn together(api_key: impl Into<String>) -> Self

Together AI: https://api.together.xyz/v1.

pub fn fireworks(api_key: impl Into<String>) -> Self

Fireworks AI: https://api.fireworks.ai/inference/v1.

pub fn perplexity(api_key: impl Into<String>) -> Self

Perplexity: https://api.perplexity.ai.

pub fn ollama() -> Self

Ollama on this machine: http://localhost:11434/v1, no key.

pub fn lm_studio() -> Self

LM Studio on this machine: http://localhost:1234/v1, no key.

pub fn with_base_url(self, base_url: impl Into<String>) -> Self

Points the client at another server. The key may be empty if the server does not need one.

pub fn with_client(self, client: Client) -> Self

Uses your own reqwest::Client, for example to share a connection pool or set a proxy.

pub fn with_max_retries(self, max_retries: u32) -> Self

How many times a failed request is retried (default DEFAULT_MAX_RETRIES; 0 disables).

Trait Implementations

impl Clone for OpenAiProvider

fn clone(&self) -> Self

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for OpenAiProvider

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

Formats the value using the given formatter.

impl EmbeddingProvider for OpenAiProvider

fn embed<'life0, 'life1, 'life2, 'life3, 'async_trait>( &'life0 self, model: &'life1 str, inputs: &'life2 [&'life3 str], ) -> Pin<Box<dyn Future<Output = Result<Embeddings>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait, 'life2: 'async_trait, 'life3: 'async_trait,

impl LlmProvider for OpenAiProvider

fn chat<'life0, 'life1, 'async_trait>( &'life0 self, req: &'life1 ChatRequest, ) -> Pin<Box<dyn Future<Output = Result<ChatResponse>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait,

Sends the request and waits for the whole answer.

fn chat_stream<'life0, 'life1, 'async_trait>( &'life0 self, req: &'life1 ChatRequest, ) -> Pin<Box<dyn Future<Output = Result<ChatStream>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait,

Streams the answer as it is generated. The last event is StreamEvent::Done with the full response.

fn list_models<'life0, 'async_trait>( &'life0 self, ) -> Pin<Box<dyn Future<Output = Result<Vec<ModelInfo>>> + Send + 'async_trait>>

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

Models available to the key.