Re-exports§
pub use anthropic::AnthropicProvider;pub use azure_openai::AzureOpenAiProvider;pub use bedrock::BedrockProvider;pub use context_translation::ContextTranslationStrategy;pub use context_translation::DefaultContextTranslation;pub use google::GoogleProvider;pub use google_vertex::GoogleVertexProvider;pub use mock::MockProvider;pub use model::ApiProtocol;pub use model::CostConfig;pub use model::CredentialProvider;pub use model::ModelConfig;pub use model::OpenAiCompat;pub use model::StaticCredentialProvider;pub use openai_compat::OpenAiCompatProvider;pub use openai_responses::OpenAiResponsesProvider;pub use registry::ProviderRegistry;pub use retry::RetryConfig;pub use traits::*;
Modules§
- anthropic
- Anthropic Claude provider (Messages API with streaming)
- azure_
openai - Azure OpenAI provider.
- bedrock
- Amazon Bedrock ConverseStream provider.
- context_
translation - G8 — Context translation strategy for cross-provider message compatibility.
- Google Generative AI (Gemini) provider.
- google_
vertex - Google Vertex AI provider.
- mock
- Mock provider for testing. No real API calls.
- model
- Model configuration and provider compatibility flags.
- openai_
compat - OpenAI Chat Completions compatible provider.
- openai_
responses - OpenAI Responses API provider.
- registry
- Provider registry — maps ApiProtocol to StreamProvider implementations.
- retry
- Retry with exponential backoff and jitter for provider calls.
- sse
- Shared SSE (Server-Sent Events) parsing utilities.
- traits