I don't like leaving the comfort of my shell >w<

A simple CLI for interacting with OpenAI's ChatGPT models.

Configuration

Default settings may be read from a configuration file:

• Linux: ~/.chatgpt-cli/.config
• Windows: C:\Users\<username>\AppData\Local\chatgpt-cli\.config

Format the file with KEY=VALUE pairs.
You may use \ to escape newlines in values,
and # at the start of a line to denote a comment.
Empty lines are ignored.

model=gpt-5-mini

# Here is a comment!
instructions=You are a helpful assistant. \ 
Respond concisely.

Environment Variables

• CHATGPT_CLI_API_KEY – Your OpenAI API key (used if not provided via --key).

Usage

./chatgpt_cli [OPTIONS] PROMPT...

Required

• -k, --key API_KEY – OpenAI API key (overrides OPENAI_API_KEY environment variable)
• -m, --model MODEL – OpenAI model to use (overrides model config option)

Optional

• -h, --help – Show help message

• -v, --version – Show program version
  
  

• -r, --raw – Print raw JSON response instead of parsed text (does not support streaming)

• -R, --response-id – Print the response id after completion (use with -H later)
  
  

• -H, --history [ID] –Specify an OpenAI previous_response_id (defaults to last response's id)

• -i, --instructions TEXT – System instructions for the model (overrides instructions config option)

• -t, --temperature DOUBLE – Sampling temperature for the model, must be in [0,2] (overrides temperature config option)

• -T, --max-tokens UINT64 – Upper bound for output tokens in the response (overrides max-tokens config option)