Skip to main content

AuthHeaderProvider

everruns_core::openai_protocol

Trait AuthHeaderProvider 

Source 
pub trait AuthHeaderProvider: Send + Sync {
    // Required method
    fn auth_header<'life0, 'async_trait>(
        &'life0 self,
    ) -> Pin<Box<dyn Future<Output = Result<(String, String)>> + Send + 'async_trait>>
       where Self: 'async_trait,
             'life0: 'async_trait;
}
Expand description

Pluggable authentication-header provider for OpenAI-compatible drivers.

When set on an OpenAIProtocolChatDriver via OpenAIProtocolChatDriver::with_auth_provider, the driver calls AuthHeaderProvider::auth_header before each request and applies the returned (name, value) header instead of the default api-key / bearer logic keyed on the host.

This lets a driver authenticate with short-lived, refreshable tokens — e.g. Microsoft Entra ID (OAuth) bearer tokens for Azure AI Foundry — without the generic protocol driver having to know the auth scheme. The provider is responsible for caching and refreshing tokens; auth_header is awaited once per HTTP attempt, so it should be cheap on the cached path.

Required Methods§

Source

fn auth_header<'life0, 'async_trait>( &'life0 self, ) -> Pin<Box<dyn Future<Output = Result<(String, String)>> + Send + 'async_trait>>
where Self: 'async_trait, 'life0: 'async_trait,

Return the (header_name, header_value) pair to apply for authentication, refreshing any cached credential as needed. Returning Err aborts the request before it is sent.

Dyn Compatibility§

This trait is dyn compatible.

In older versions of Rust, dyn compatibility was called "object safety".

Implementors§