Enum Provider 

#[non_exhaustive]
pub enum Provider {
    OpenAI {
        api_key: String,
        api_base: String,
        model: String,
    },
    Ollama {
        base_url: String,
        model: String,
    },
}

LLM Provider configuration

Each variant is feature-gated to ensure only enabled providers are available. Use Provider::from_env() to automatically select based on environment variables.

Variants (Non-exhaustive)

Non-exhaustive enums could have additional variants added in future. Therefore, when matching against variants of non-exhaustive enums, an extra wildcard arm must be added to account for any future variants.

OpenAI

Available on crate feature openai only.

OpenAI API and compatible endpoints (e.g., Azure OpenAI, local vLLM)

Fields

api_key: String

API key for authentication

api_base: String

Base URL for the API (default: https://api.openai.com/v1)

model: String

Model identifier (e.g., “gpt-4”, “gpt-3.5-turbo”)

Ollama

Available on crate feature ollama only.

Ollama local inference server

Fields

base_url: String

Base URL for Ollama API (default: http://localhost:11434)

model: String

Model name (e.g., “ministral-3:3b”, “mistral”, “qwen3-vl:2b”)

Implementations

impl Provider

pub async fn create_client(&self) -> Result<Box<dyn LLMClient>>

Create an LLM client from this provider configuration

Errors

Returns an error if:

• The provider cannot be initialized
• Required configuration is missing
• Network connectivity issues (for remote providers)

pub fn from_env() -> Result<Self>

Create a provider from environment variables

Provider priority (first match wins):

1. LlamaCpp - if LLAMACPP_MODEL_PATH is set
2. OpenAI - if OPENAI_API_KEY is set
3. Ollama - default fallback for local inference

Environment Variables

LlamaCpp

• LLAMACPP_MODEL_PATH - Path to GGUF model file (required)

OpenAI

• OPENAI_API_KEY - API key (required)
• OPENAI_API_BASE - Base URL (default: https://api.openai.com/v1)
• OPENAI_MODEL - Model name (default: gpt-4)

Ollama

• OLLAMA_BASE_URL - Server URL (default: http://localhost:11434)
• OLLAMA_MODEL - Model name (default: ministral-3:3b)

Errors

Returns an error if no LLM provider features are enabled or configured.

Example

// Set environment variables
std::env::set_var("OLLAMA_MODEL", "ministral-3:3b");

let provider = Provider::from_env()?;
let client = provider.create_client().await?;

pub fn name(&self) -> &'static str

Get the provider name as a string

pub fn requires_api_key(&self) -> bool

Check if this provider requires an API key

pub fn is_local(&self) -> bool

Check if this provider is local (no network required)

pub fn from_config( provider_config: &ProviderConfig, model_override: Option<&str>, ) -> Result<Self>

Create a provider from TOML configuration

Arguments

• provider_config - The provider configuration from ares.toml
• model_override - Optional model name to override the provider default

Errors

Returns an error if the provider type doesn’t match an enabled feature or if required environment variables are not set.

pub fn from_model_config( model_config: &ModelConfig, provider_config: &ProviderConfig, ) -> Result<Self>

Create a provider from a model configuration and its associated provider config

This is the primary way to create providers from TOML config, as it resolves the model -> provider reference chain.

Trait Implementations

impl Clone for Provider

fn clone(&self) -> Provider

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for Provider

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

Formats the value using the given formatter.