CliGPT

Description

A bash script that use openai api to generate text, image and speech.

Required Dependencies

general dependencies

• jq
• bat
• curl
• xsel
• notify-send

speech dependencies

• pacat
• zenity

tts dependencies

• paplay

Installation

1. Clone the repository:

git clone https://github.com/LordPax/cligpt.git
cd cligpt

2. Execute the script to generate config file in .config/cligpt

./cligpt

3. Add your API key to ~/.config/cligpt/config:

Example

API_KEY=your-api-key
API_ROUTE=https://api.openai.com/v1
TEXT_MODEL=gpt-4
TEXT_TEMP=0.7
IMAGE_MODEL=dall-e-3
IMAGE_NB=1
IMAGE_SIZE=1024x1024
SPEECH_MODEL=whisper-1
SPEECH_LANG=fr
TTS_MODEL=tts-1-hd
TTS_VOICE=alloy
TTS_SPEED=1.0
TTS_FORMAT=mp3

Integration with tmux

Add the following line to your .tmux.conf:

bind H new-window "cligpt text" \; rename-window "cligpt"

Vim plugin

You can find the vim plugin here

Add the following line to your .vimrc:

Plug 'LordPax/vim-cligpt'

Integration with i3

Add the following line to your ~/.config/i3/config:

bindsym $mod+s exec ~/path/to/cligpt -c -g speech

Usage

• main help

Usage : cligpt [option] [command]

Option :
-h or --help ........................... Show help
-v or --version ........................ Show version
-l or --list ........................... List all available model
-i or --inerte ......................... Do nothing except saving prompt in history
-q or --quiet .......................... Quiet mode
-c or --clipboard ...................... Save result in clipboard
-g or --gui ............................ Open GUI

Command :
text ................................... Generate text
image .................................. Generate image
speech ................................. Generate text from speech
tts .................................... Generate speech from text

• text help

Usage : cligpt [generalOption] text [option] [prompt|-]

Option :
-h or --help ........................... Show help text command
-m or --model <model> .................. Select model (default : gpt-4)
-t or --temp <temperature> ............. Set temperature (default : 0.7)
-s or --system [text] .................. Instruction with role system (use "-" for stdin)
-c or --clear .......................... Clear history
-l or --list-history ................... List history
--list-json-history .................... List history in json format

Example :
cligpt text # Interactive mode
cligpt text "some prompt with role user"
echo "some prompt with role user" | cligpt text -
cat somefile.txt | cligpt text -s - # Instruction mode
cat somefile.txt | cligpt text -s - "some prompt with role user"
cligpt text -s "some prompt with role system" "some prompt with role user"

• image help

Usage : cligpt [generalOption] image [option] <prompt|->

Option :
-h or --help ........................... Show help image command
-m or --model <model> .................. Select model (default : dall-e-3)
-n or --nb <number> .................... Number of image to generate (default : 1)
-s or --size <size> .................... Size of image (default : 1024x1024)
-o or --output <name> .................. Get image in file

Size :
- 256x256
- 512x512
- 1024x1024
- 1024x1792
- 1792x1024

Example :
cligpt image "some prompt"
echo "some prompt" | cligpt image -
cligpt image -o image.png "some prompt"

• speech help

Usage : cligpt [generalOption] speech [option] [file]

Option :
-h or --help ........................... Show help speech command
-m or --model <model> .................. Select model (default : whisper-1)
-l or --lang <language> ................ Select language (default : fr)

Example :
cligpt speech # record audio and generate text
cligpt speech somefile.ogg
cligpt -c -g speech

• tts help

Usage : cligpt [generalOption] tts [option] [text|-]

Option :
-h or --help ........................... Show help tts command
-m or --model <model> .................. Select model (default : tts-1-hd)
-v or --voice <voice> .................. Select voice (default : alloy)
-s or --speed <speed> .................. Select speed (default : 1.0)
-f or --format <format> ................ Select format (default : mp3)
-o or --output <name> .................. Get speech in file

Voice :
- alloy
- echo
- fable
- onyx
- nova
- shimmer

Example :
cligpt tts "some text" # play audio
echo "some text" | cligpt tts -
cligpt tts -o speech.mp3 "some text"

FAQ

How do I get my OpenAI API key?

To get an OpenAI API key, you need to sign up on the OpenAI website and follow the instructions to create an account and obtain an API key.

Example

TODO : update example