🏁 Runtime Command-Line Options

This page documents the command-line options that temporarily override configuration for a single session (do not persist changes to config files).

💡 Overview

These options are useful for one-off runs, scripting, or experimentation. They take precedence over config files for the current invocation only.

Set the maximum number of tokens for the model response.

janito "Your prompt here" --max-tokens 500

Limit the number of tool calls allowed within a chat session.

janito "Prompt" --max-tools 3

Specify the model name for this session.

janito "Prompt" --model openai/gpt-4.1

Set the maximum number of agent rounds per prompt.

janito "Prompt" --max-rounds 10

Override the system prompt or role for this session.

janito "Prompt" --system "You are a helpful assistant."
janito "Prompt" --system-file my_prompt.txt
janito "Prompt" --role "expert Python developer"

Set the sampling temperature for the model.

janito "Prompt" --temperature 0.7

Control tool usage and output for the session.

janito "Prompt" --no-tools
janito "Prompt" --vanilla
janito "Prompt" --trust-tools

Set the interaction style for the system prompt template.

janito "Prompt" --style technical

Enable streaming output from OpenAI and print raw chunks.

janito "Prompt" --stream
janito "Prompt" --verbose-stream

Set a runtime (in-memory only) config key-value pair. Can be repeated.

janito "Prompt" --run-config key1=val1 --run-config key2=val2