llm/

Command advanced shows two lower-level controls layered on a normal request:

Command basic sends a single prompt to a model and prints the reply.

Command conversation carries history across multiple turns.

Command model_switch continues one conversation across two protocols.

Command options sends a prompt with a system message and per-request options.

Command reasoning asks a reasoning-capable model to think before answering and streams the reasoning and the final answer as separate phases.

Command streaming consumes a response as a live event stream instead of waiting for the final message.

Command tools runs a streaming tool loop with reasoning: the model thinks, optionally calls a typed tool, sees the result, and continues until it gives a final answer.