Crate runtara_ai 

AI/LLM integration for runtara workflows.

This crate provides a synchronous completion abstraction:

• CompletionModel trait and request builder
• Message types (user, assistant, tool calls, tool results)
• OneOrMany<T> non-empty collection
• Provider implementations (OpenAI-compatible)
• Provider dispatch (connection → CompletionModel)
• Shared types for conversation history, tool call logs, and usage tracking

Re-exports

pub use completion::CompletionError;

pub use completion::CompletionModel;

pub use completion::CompletionRequest;

pub use completion::CompletionResponse;

pub use message::AssistantContent;

pub use message::Message;

pub use message::UserContent;

pub use one_or_many::OneOrMany;

Modules

completion

Synchronous completion model abstraction.

message

Message types for AI Agent conversations.

one_or_many

OneOrMany<T> — a non-empty collection that holds one or more items.

provider

LLM provider dispatch.

providers

LLM provider implementations.

types

Shared types for AI Agent execution.