OptionsOn gptme itself
--model-m
For example anthropic/claude-sonnet-4-5 or just a provider name. <model>
--workspace-w
Workspace directory <workspace>
--resume-r
Resume most recent conversation
--name
Conversation name <name>
--non-interactive-n
Run the prompt to completion without asking questions; implies skipping confirmations.
--no-confirm-y
Lets tools run without asking first.
--output-format
Output format (non-interactive) <output_format>
--tools-t
Comma-separated tool names to allow. <tools>
--no-workspace
Skip workspace context
