SGPT

SGPT (aka shell-gpt) is a powerful command-line interface (CLI) tool designed for seamless interaction with OpenAI models directly from your terminal. Effortlessly run queries, generate shell commands or code, create images from text, and more, using simple commands. Streamline your workflow and enhance productivity with this powerful and user-friendly CLI tool.

Developed with the help of SGPT.

This is a Go implementation. For the original Python implementation, visit shell-gpt. Please keep this in mind when reporting issues.

• Features
• Installation
  • Linux
  • macOS
  • Windows
  • Using Go
  • Docker
  • Ansible
  • Other platforms
• Usage Guide
  • Getting started: Obtaining an OpenAI API Key
  • Querying OpenAI Models
  • GPT-4o and GPT-4 Vision API Support
  • o1 API Support
  • OpenRouter API Support
  • Requesty API Support
  • Local LLM Support
  • Chat Capabilities
  • Generating and Executing Shell Commands
  • Interactive Shell Sessions
  • Code Generation Capabilities
  • Prompt Templating
  • Enhancing Your Workflow with Bash Aliases and Functions
• Acknowledgements

Features

• Instant Answers: Obtain quick and accurate responses to simple questions directly in your shell, streamlining your workflow.
• GPT-4o Integration: Access the capabilities of the GPT-4o API to generate detailed and informative responses.
• GPT-4 Vision API: Leverage the capabilities of the GPT-4 Vision API to analyze and generate insights from images.
• Shell Commands Generation: Effortlessly generate and execute shell commands, simplifying complex tasks and enhancing productivity.
• Code Production: Generate code snippets in various programming languages, making it easier to learn new languages or find solutions to coding problems.
• ChatGPT Integration: Utilize ChatGPT's interactive chat capabilities to refine your prompts and achieve more precise results, benefiting from the powerful language model.
• Bash Functions and Aliases: Seamlessly integrate SGPT responses into custom bash functions and aliases, optimizing your workflows and making your daily tasks more efficient.
• OpenRouter Support: Use OpenRouter to access various large language models (LLMs) via a single API, providing flexibility and convenience in your interactions with different models.
• Prompt Templating: Inject structured data into prompts using Go text/template syntax. Pipe YAML or JSON variables via stdin to build dynamic, reusable prompt patterns.

By offering these versatile features, SGPT serves as a powerful tool to enhance your overall productivity, streamline your workflow, and simplify complex tasks.

Installation

Linux

SGPT has been tested on Ubuntu LTS releases and is expected to be compatible with the following Linux distributions:

• Debian
• Ubuntu
• Arch Linux
• Fedora

To install, download the latest release from the release page and use the package manager specific to your distribution.

macOS

For users with Homebrew as their package manager, run the following command in the terminal:

brew install tbckr/tap/sgpt

Windows

For users with Scoop as their package manager, execute these commands in PowerShell:

scoop bucket add tbckr https://github.com/tbckr/scoop-bucket.git
scoop install tbckr/sgpt

Using Go

To install SGPT with Go, based on the git tag, use this command:

go install github.com/tbckr/sgpt/v2/cmd/sgpt@latest

Docker

To run SGPT with Docker, use the following command to pull the latest image:

docker pull ghcr.io/tbckr/sgpt:latest

Examples on how to use SGPT with Docker can be found here.

Ansible

To install SGPT with Ansible, you can use the following ansible playbook as your base and adapt accordingly:

---
- hosts: all
  tasks:
    - name: Get latest sgpt release
      uri:
        url: "https://api.github.com/repos/tbckr/sgpt/releases/latest"
        return_content: yes
      register: sgpt_release

    - name: Set latest version of sgpt
      set_fact:
        sgpt_latest_version: "{{ sgpt_release.json.tag_name }}"

    - name: Install sgpt for debian based, amd64 systems
      ansible.builtin.apt:
        deb: https://github.com/tbckr/sgpt/releases/download/{{ sgpt_latest_version }}/sgpt_{{ sgpt_latest_version[1:] }}_amd64.deb
        allow_unauthenticated: true

The playbook can be run with the following command:

ansible-playbook -i <inventory> <playbook>.yml

The latest version of the playbook can be found here.

Other platforms

For other platforms, visit the GitHub release page and download the latest release suitable for your system.

Usage Guide

See the documentation for detailed usage instructions.

Getting started: Obtaining an OpenAI API Key

To use the OpenAI API, you must first obtain an API key.

1. Visit https://platform.openai.com/overview and sign up for an account.
2. Navigate to https://platform.openai.com/account/api-keys and generate a new API key.
3. On Linux or macOS: Update your .bashrc or .zshrc file to include the following export statement adding your API key as the value:

export OPENAI_API_KEY="sk-..."

1. On Windows: Update your environment variables to include the OPENAI_API_KEY variable with your API key as the value.

After completing these steps, you'll have an OpenAI API key that can be used to interact with the OpenAI models through the SGPT tool.

[!IMPORTANT] Your API key is sensitive information. Do not share it with anyone.

Querying OpenAI Models

SGPT allows you to ask simple questions and receive informative answers. For example:

$ sgpt "mass of sun"
The mass of the sun is approximately 1.989 x 10^30 kilograms.

You can also pass prompts to SGPT using pipes:

$ echo -n "mass of sun" | sgpt
The mass of the sun is approximately 1.989 x 10^30 kilograms.

You can also add another prompt to the piped data by specifying the stdin modifier and then specifying the prompt:

$ echo "Say: Hello World!" | sgpt stdin 'Replace every "World" word with "ChatGPT"'
Hello ChatGPT!

If you want to stream the completion to the command line, you can add the --stream flag. This will stream the output to the command line as it is generated.

GPT-4o and GPT-4 Vision API Support

SGPT additionally facilitates the utilization of the GPT-4o and GPT-4 Vision API. Include input images using the -i or --input flag, supporting both URLs and local images.

$ sgpt -m "gpt-4o" -i pkg/fs/testdata/marvin.jpg "what can you see on the picture?"
The picture shows a robot with a large, round head and an expressive, downward-slanting triangular eye. The body of the robot is designed with a sleek, somewhat shiny, metallic structure and it is pointing with its right hand. The design appears to be humanoid with distinct arms, legs, and a segmented torso.
$ sgpt -m "gpt-4-vision-preview" -i "https://upload.wikimedia.org/wikipedia/en/c/cb/Marvin_%28HHGG%29.jpg" "what can you see on the picture?"
The image shows a figure resembling a robot with a humanoid form. It has a

It is also possible to combine URLs and local images:

$ sgpt -m "gpt-4o" -i "https://upload.wikimedia.org/wikipedia/en/c/cb/Marvin_%28HHGG%29.jpg" -i pkg/fs/testdata/marvin.jpg "what is the difference between those two pictures"
The two pictures you provided appear to be identical. There are no visible differences between them. Both show the same character in the same pose with the same lighting and background.

You can also set the default model to GPT-4o or GPT-4 Vision by setting it in the configuration file.

Important: The GPT-4o and GPT-4-vision API integration is currently in beta and may change in the future.

o1 API Support

If you are already whitelisted for the o1 API, you can use it by specifying the model with the -m flag. You must also provide the --stream=false flag to not stream the output as it is not supported by the o1 API (this is only necessary, if you have provided the stream option via the config file).

Example:

$ sgpt -m "o1-preview" --stream=false "how many rs are in strawberry?"
There are three "r"s in the word "strawberry".

You can also create a bash alias to use the o1 API more easily. For example, add the following line to your .bashrc:

alias sgpt-o1="sgpt -m \"o1-preview\" --stream=false"

Then you can use the alias like this:

$ sgpt-o1 "how many rs are in strawberry?"
There are three "r"s in the word "strawberry".

Important: The o1 API does not support personas.

OpenRouter API Support

SGPT seamlessly integrates with the OpenRouter API, giving you access to a wide range of AI models beyond OpenAI's offerings.

1. Set the OpenRouter API base URL environment variable:

   export OPENAI_API_BASE="https://openrouter.ai/api/v1"
   

2. Create an API key at OpenRouter and set it as your environment variable:

   export OPENAI_API_KEY="your_openrouter_api_key"
   

Once configured, you can specify any OpenRouter-supported model with the -m flag:

$ sgpt -m "