Module llm_agent 

LLM provider abstraction.

LlmAgent wraps multiple large-language-model backends behind a single async complete() interface. Supported providers:

• Anthropic (claude-* models, Messages API)
• OpenAI (gpt-* and compatible, Chat Completions API)
• Ollama (local, OpenAI-compatible endpoint)

The active provider and model are selected via LlmConfig.

Structs

ChatMessage

A single turn in a conversation (role + content).

LlmAgent

An actor that calls an LLM provider and returns completions.

LlmConfig

Configuration for the LLM backend.

Enums

LlmProvider

Supported LLM provider backends.

Functions

calc_cost_nano_usd

Calculate cost in nano-USD from token counts and model name.