ChatGPT-cli

A way to communicate with ChatGPT from the command line.

Installing

brew tap duanemay/tap
brew install chatgpt-cli

Running

API Key

To utilize the ChatGPT CLI, a ChatGPT API key is required. You can obtain one by signing up and creating an API key at API keys.

Flags, Environment Variables

Configuration settings follow the precedence order: flags > environment variables > configuration files > defaults

Long	Short	Config File/Environment Variable	Default	Description
`--api-key`	`-k`	`CHATGPT_API_KEY`	Error	ChatGPT API Key
`--config`	`-c`	`CHATGPT_CONFIG`	./.chatgpt-cli then $HOME/.chatgpt-cli	Config file to load
`--verbose`	`-v`	`CHATGPT_VERBOSE`	`false`	Verbose logging
`--eom`		`CHATGPT_EOM`	`\s`	End of message marker
`--eos`		`CHATGPT_EOS`	`\q`	End of session marker
`--session-file`	`-s`	`CHATGPT_SESSION_FILE`	Generated	Session file
`--model`	`-m`	`CHATGPT_MODEL`	`gpt-4`	Model to use
`--role`		`CHATGPT_ROLE`	`user`	Role of User
`--temperature`	`-t`	`CHATGPT_TEMPERATURE`	`1.0`	Temperature: 0-2
`--max-tokens`		`CHATGPT_MAX_TOKENS`	`16`	Max tokens
`--top-p`		`CHATGPT_TOP_P`	`1.0`	Top P: 0-1

chatgpt-cli is a CLI for ChatGPT

Usage:
  chatgpt-cli [command]

Available Commands:
  chat          Enter a chat session with ChatGPT
  completion    Generate the autocompletion script for the specified shell
  help          Help about any command
  listModels    lists all models available to your account
  replaySession Replay a chat session from saved file
  version       displays version information

Flags:
  -k, --api-key string   ChatGPT apiKey
  -c, --config string    Config file (default ./.chatgpt-cli then $HOME/.chatgpt-cli)
  -h, --help             help for chatgpt-cli
  -v, --verbose          verbose logging

Use "chatgpt-cli [command] --help" for more information about a command.

Commands

chat

Enter a chat session with ChatGPT

chatgpt-cli chat

Help for chat command

Enter a chat session with ChatGPT

Usage:
  chatgpt-cli chat [flags]

Flags:
      --eom string            Text to enter to mark the end of a message to send to ChatGPT (default "\s")
      --eos string            Text to enter to end of a session with ChatGPT (default "\q")
  -h, --help                  help for chat
      --max-tokens int        number of tokens to generate = $ (default 9223372036854775807)
  -m, --model string          ChatGPT Model (default "gpt-4")
  -r, --role string           ChatGPT Role (default "user")
  -s, --session-file string   Continue a session from a file
  -t, --temperature float32   temperature, between 0 and 2. Higher values make the output more random (default 1)
      --top-p float32         results of the tokens with top_p probability mass (default 1)

Global Flags:
  -k, --api-key string   ChatGPT apiKey
  -c, --config string    Config file (default ./.chatgpt-cli then $HOME/.chatgpt-cli)
  -v, --verbose          verbose logging

This will enter a chat session with ChatGPT. You will be prompted for a message. Enter a message you can enter multi-line text to send. When your message is complete enter CTRL+D or \s on its own line to send. (You can change the default end of message marker with the --eom flag)

ChatGPT will respond with a message. You can continue the conversation by entering another message.

To exit the chat session, enter CTRL+C or \q on its own line. Any text already entered will not be sent. (You can change the default end of session marker with the --eos flag)

Your session will be saved in session file. The file name will be displayed when you enter the chat session. You can resume the session by specifying the file using the --session-file flag.

chatgpt-cli chat --session-file session.json

Model, Role, Temperature, Max Tokens, Top P can be set with the --model, --role, --temperature, --max-tokens, --top-p flags. These can be changed when resuming a session with the --session-file flag.

replaySession

Replay a chat session from saved file, this is useful for displaying a chat session in a easier to read format, than the raw JSON.

chatgpt-cli replaySession --session-file session.json

Help for replaySession command

Usage:
  chatgpt-cli replaySession [flags]

Flags:
  -h, --help                 help for replaySession
  -s, --sessionFile string   File to replay a Session from

Global Flags:
  -k, --api-key string   ChatGPT apiKey
  -c, --config string    Config file (default ./.chatgpt-cli then $HOME/.chatgpt-cli)
  -v, --verbose          verbose logging

listModels

lists all models available to your account

chatgpt-cli listModels

Help for listModels command

 lists all models available to your account

Usage:
  chatgpt-cli listModels [flags]

Flags:
  -h, --help   help for listModels

Global Flags:
  -k, --api-key string   ChatGPT apiKey
  -c, --config string    Config file (default ./.chatgpt-cli then $HOME/.chatgpt-cli)
  -v, --verbose          verbose logging

version

displays version information

chatgpt-cli version

Build & Release

Do a trial run of the deployment process. This is as easy as running a command with a few flags that will stop you from releasing to GitHub.

goreleaser --snapshot --skip-publish --clean

After adding the GitHub token in ./.github_token Run:

goreleaser --clean

# README