Crate brainwires_provider 

Provider layer for the Brainwires Agent Framework.

Contains both low-level API client structs (HTTP transport, auth, rate limiting, serde) and high-level chat provider implementations that wrap them with the brainwires_core::Provider trait.

Re-exports

pub use http_client::RateLimitedClient;

pub use rate_limiter::RateLimiter;

pub use brainwires_http::DEFAULT_BACKEND_URL;

pub use brainwires_http::DEV_BACKEND_URL;

pub use brainwires_http::get_backend_from_api_key;

pub use anthropic::AnthropicClient;

pub use brainwires_http::BrainwiresHttpProvider;

pub use gemini::GoogleClient;

pub use ollama::OllamaProvider;

pub use openai_chat::OpenAiClient;

pub use anthropic::chat::AnthropicChatProvider;

pub use gemini::chat::GoogleChatProvider;

pub use ollama::chat::OllamaChatProvider;

pub use openai_chat::chat::OpenAiChatProvider;

pub use openai_responses::OpenAiResponsesProvider;

pub use model_listing::AvailableModel;

pub use model_listing::ModelCapability;

pub use model_listing::ModelLister;

pub use model_listing::create_model_lister;

pub use chat_factory::ChatProviderFactory;

pub use local_llm::*;

Modules

anthropic

Anthropic Messages protocol (also used by Bedrock, Vertex AI).

brainwires_http

Brainwires HTTP relay protocol.

chat_factory

Chat provider factory — registry-driven protocol dispatch. Chat provider factory — registry-driven protocol dispatch.

gemini

Google Gemini generateContent protocol.

http_client

Rate-limited HTTP client wrapper.

local_llm

Local LLM inference (always compiled, llama.cpp behind feature flag). Local LLM Provider Module

model_listing

Model listing — query available models from provider APIs. Model listing and validation for AI providers.

ollama

Ollama native chat protocol.

openai_chat

OpenAI Chat Completions protocol (also used by Groq, Together, Fireworks, Anyscale). OpenAI (and OpenAI-compatible) API client, wire types, and submodules.

openai_responses

OpenAI Responses API protocol (/v1/responses). OpenAI Responses API provider (POST /v1/responses).

rate_limiter

Token-bucket rate limiter for API request throttling.

registry

Provider registry — protocol, auth, and endpoint metadata for all known providers. Provider registry — connection details for all known providers.

Structs

ChatOptions

Chat completion options

ProviderConfig

Provider configuration

Enums

ProviderType

AI provider types

Traits

Provider

Base provider trait for AI providers