Crate cognate_providers 

Cognate Providers — concrete LLM provider implementations.

Providers

Provider	Struct	Chat	Stream	Tools	Embeddings
OpenAI	OpenAiProvider	✓	✓	✓	✓
Anthropic	AnthropicProvider	✓	✓	✓	—

Resilience

• All providers have built-in exponential-backoff retry via RetryConfig.
• Token-bucket rate limiting is available on OpenAiProvider and AnthropicProvider.
• FallbackProvider transparently retries with a secondary provider on any retryable error.

Re-exports

pub use anthropic::AnthropicProvider;

pub use costs::ModelCost;

pub use fallback::FallbackProvider;

pub use openai::OpenAiProvider;

pub use retry::with_retry;

pub use retry::RetryConfig;

pub use sse::SseStream;

Modules

anthropic

Anthropic provider implementation.

costs

Model pricing constants.

fallback

Transparent provider fallback.

openai

OpenAI provider implementation.

retry

Retry logic with exponential backoff

sse

Server-Sent Events (SSE) streaming parser

Constants

DEFAULT_TIMEOUT

Default request timeout applied when no explicit timeout is configured.

Functions

create_http_client

Build a reqwest client with standard Cognate defaults.