Skip to content

Usage

Command Syntax

chatgpt-cli <command> [arguments]

Running chatgpt-cli without any arguments displays the help message.

Available Commands

Command Description
help Show help message with all available commands
prompt <text> Send a prompt to ChatGPT and display the response
logs Display application logs
config <subcommand> Manage configuration (list, get, set)

help

Displays usage information, all available commands, configuration options, and examples.

Syntax:

chatgpt-cli help

This is also the default command when no arguments are provided.

prompt

Sends a prompt to ChatGPT using the OpenAI API and prints the response to standard output.

Syntax:

chatgpt-cli prompt <text>

Arguments:

Argument Required Description
text Yes The prompt text to send to ChatGPT. Can be a quoted string or multiple words.

Behavior:

  • Requires the OPENAI_API_KEY to be set (via environment variable or config file).
  • Multiple arguments are joined with spaces to form the prompt.
  • The prompt cannot be empty or whitespace-only.
  • Successful interactions are logged to the log file.
  • Failed interactions are also logged with the error message.

Examples:

# Ask a simple question
chatgpt-cli prompt "What is Go?"

# Ask for code generation
chatgpt-cli prompt "Write a function to reverse a string in Go"

# Multi-word prompt without quotes
chatgpt-cli prompt Explain async await in JavaScript

# Override model for a single query using an environment variable
OPENAI_MODEL=gpt-4 chatgpt-cli prompt "Explain quantum computing"

# Save response to a file
chatgpt-cli prompt "Write a README for my project" > output.md

# Use in a script
RESPONSE=$(chatgpt-cli prompt "Generate a git commit message")
git commit -m "$RESPONSE"

Errors:

Error Cause
prompt text is required No prompt text was provided
prompt cannot be empty The prompt text is empty or whitespace-only
missing API key OPENAI_API_KEY is not set
failed to get response API request failed (network error, timeout, invalid key, etc.)

logs

Displays all stored application logs, including timestamps, prompts, responses, and errors.

Syntax:

chatgpt-cli logs

Logs are stored in JSONL format at <config_dir>/logs.jsonl. Each log entry contains:

  • Timestamp — when the event occurred
  • Command — the command that was executed
  • Prompt — the prompt text (if applicable)
  • Response — the ChatGPT response (if applicable)
  • Error — the error message (if the command failed)

Long prompts and responses are truncated to 80 characters in the display output.

Example Output:

Showing 2 log entries:

[1] 2024-01-31 14:30:15 - prompt
    Prompt: What is Go?
    Response: Go is a statically typed, compiled programming language...

[2] 2024-01-31 14:35:22 - prompt
    Prompt: Explain channels
    Error: failed to get response: unexpected status code: 401

config

Manages application configuration. Has three subcommands: list, get, and set.

Syntax:

chatgpt-cli config <subcommand> [arguments]

config list

Lists all current configuration values. The API key is masked for security (shows first 4 and last 4 characters, or *** for short keys).

chatgpt-cli config list

Example Output:

Current Configuration:
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
OPENAI_API_KEY:           sk-a...b1c2
OPENAI_API_URL:           https://api.openai.com/v1/chat/completions
OPENAI_MODEL:             gpt-3.5-turbo
OPENAI_TIMEOUT:           1m0s
OPENAI_MAX_TOKENS:        1000
OPENAI_TEMPERATURE:       0.7
CHATGPT_CLI_CONFIG_DIR:   /home/user/.chatgpt-cli

config get

Gets the current value of a specific configuration key. The key name is case-insensitive.

chatgpt-cli config get <key>

Valid keys: OPENAI_API_KEY, OPENAI_API_URL, OPENAI_MODEL, OPENAI_TIMEOUT, OPENAI_MAX_TOKENS, OPENAI_TEMPERATURE, CHATGPT_CLI_CONFIG_DIR

Examples:

chatgpt-cli config get OPENAI_MODEL
# Output: gpt-3.5-turbo

chatgpt-cli config get openai_max_tokens
# Output: 1000

# Use in a script
MODEL=$(chatgpt-cli config get OPENAI_MODEL)
echo "Current model: $MODEL"

config set

Sets a configuration value and persists it to the config file.

chatgpt-cli config set <key> <value>

Settable keys and validation rules:

Key Validation
OPENAI_API_KEY Cannot be empty
OPENAI_API_URL Must start with http:// or https://
OPENAI_MODEL Cannot be empty
OPENAI_TIMEOUT Must be a valid Go duration (e.g., 60s, 1m, 90s)
OPENAI_MAX_TOKENS Must be a positive integer
OPENAI_TEMPERATURE Must be a number between 0.0 and 2.0

!!! note CHATGPT_CLI_CONFIG_DIR cannot be set via config set. Use the environment variable instead.

Examples:

# Switch to GPT-4
chatgpt-cli config set OPENAI_MODEL gpt-4

# Increase max tokens
chatgpt-cli config set OPENAI_MAX_TOKENS 2000

# Set a longer timeout
chatgpt-cli config set OPENAI_TIMEOUT 120s

# Adjust creativity
chatgpt-cli config set OPENAI_TEMPERATURE 1.5

# Set a custom API endpoint
chatgpt-cli config set OPENAI_API_URL https://my-proxy.example.com/v1/chat/completions

Successful output:

Set OPENAI_MODEL=gpt-4
Configuration saved to /home/user/.chatgpt-cli/config

Unknown Commands

If you provide an unrecognized command, the CLI prints an error message and displays the help text:

$ chatgpt-cli foo
Unknown command: foo

ChatGPT CLI - Command Line Interface for ChatGPT
...

For configuration details, see the Configuration page.