Google Gemini-AI CLI App

GeminiAI-Terminal is a powerful command-line interface (CLI) application that leverages Google's Gemini AI to generate text based on user prompts. This allows users to interact with advanced AI capabilities directly from their terminal, making it an essential tool for developers, and AI enthusiasts who prefer a terminal workflow.

Table of Contents

• Features

• Prerequisites

• Installation

  • Execute the Binary Directly - Quick & Easy method
  • Compile and Build the Executable

• Configuration

  • Configuration Explanation
  • Safety Settings

• Usage

• Command-Line Arguments

• Examples

• License

• Author

Features

• Text-to-Text Model - Engage in conversations with Gemini-AI using text Prompts
• Multi-Mode Model - Get output from given inputs text prompt and image
• Clipboard Integration - Instead of manually typing the file path, copy path to the clipboard and press enter
• Flexible Customizations - Flexibly configure/modify configurations from a single JSON file
• Colorful Engaging Terminal Outputs - Enjoy colorful engaging outputs including Code highlighting

Prerequisites

• Gemini API Key To use this app, you'll need an API key. If you don't already have one, create a key in Google AI Studio Create an API key from Google AI Studio

• Go version 1.20 or higher

• For best experience, use Gemini AI with Yakuake terminal to feel the productivity boost (optional)

  sudo pacman -S yakuake

Installation

Run the Binary (Quick and easy method)

Download the app from GitHub Release, Extract the ZIP file and use directly without building/compiling.

Use the app directly without Compiling or downloading dependencies,

Download latest version of the app from GitHub Release, Extract the ZIP file and Execute as shown below.

For Linux users

# first navigate to the directory where the zip file is
unzip Gemini-AI-CLI-App.zip
chmod +x gemini
./gemini

You have to setup the API key As shown here

Follow these steps on topic #Usage Method 3, to use gemini as a regular Linux command.

! Only tested on Linux

---

Compile/Build the program

1. Clone the Repository

   git clone https://github.com/mr-alham/Google-gemini-AI-CLI-App.git
   cd Google-gemini-AI-CLI-App

2. Install Dependencies

   • Go Generative-AI SDK Package

     go get github.com/google/generative-ai-go

   • Options to configure Google API client

     go get google.golang.org/api/option

   • Clipboard

     go get github.com/atotto/clipboard

   • Glamour

     go get github.com/charmbracelet/glamour

   • Utilities to Interact with Terminal

     go get golang.org/x/term

3. Build the Application

   go build -o gemini

Configuration

The application can be configured with a single configuration file keys.json. Below is the configuration file with explanations for each configuration,

{
    "GEMINI_API_KEY": "your gemini api key",
    "GEMINI_MODEL": "gemini-1.5-pro-latest",
    "SYSTEM_INSTRUCTION": "",
    "GENERATION_CONFIG": {
        "Temperature": 0.9,
        "top_p": 0.95,
        "top_k": 100,
        "max_output_tokens": 8192,
        "response_mime_type": "text/plain"
    },
    "SAFETY_SETTINGS": [
        {
            "category": "HARM_CATEGORY_HARASSMENT",
            "threshold": "BLOCK_MEDIUM_AND_ABOVE"
        },
        {
            "category": "HARM_CATEGORY_HATE_SPEECH",
            "threshold": "BLOCK_MEDIUM_AND_ABOVE"
        },
        {
            "category": "HARM_CATEGORY_SEXUALLY_EXPLICIT",
            "threshold": "BLOCK_MEDIUM_AND_ABOVE"
        },
        {
            "category": "HARM_CATEGORY_DANGEROUS_CONTENT",
            "threshold": "BLOCK_MEDIUM_AND_ABOVE"
        }
    ]
}

Configuration Explanation

• GEMINI_API_KEY

  • Your API key for accessing the Gemini AI services. This is required to authenticate your requests. Replace your gemini api key with your API key

    { "GEMINI_API_KEY": "your gemini api key" }

• GEMINI_Model

  • Generative AI models are able to create content from varying types of data input including text, images and audio.

  Model	Input	Optimized for
  gemini-1.5-pro (Default)	Audio, Image,Video and Text	Complex reasoning tasks such as code and text generation, text editing, problem solving, data extraction and generation
  gemini-1.5-flash	Audio, Image,Video and Text	Fast and versatile performance across a diverse variety of tasks
  gemini-1.0-pro	Text	Natural language tasks, multi-turn text and code chat, and code generation
  gemini-pro-vision	Audio, Image,Video and Text	Visual-related tasks, like generating image descriptions or identifying objects in images

• SYSTEM_INSTRUCTION

  • System instructions enable users to steer the behavior of the model based on their specific needs and use cases

    "SYSTEM_INSTRUCTION": "You are a cat. Your name is Neko."

• Temperature

  • The temperature controls the degree of randomness in token selection The temperature can be changed in the range of 0 to 1

• top_p

  • model alters the way tokens are selected for output. It involves selecting tokens from the most probable to the least probable until their cumulative probability equals the topP value. For example, if tokens A, B, and C have probabilities of 0.3, 0.2, and 0.1, and the topP value is 0.5, the model picks either A or B as the next token using temperature sampling, excluding C. The default topP value is 0.95. This parameter helps filter the most probable tokens and exclude less probable ones.

• top_K

  • determines the number of the most probable tokens to be considered for output selection. A topK value of 1 (greedy decoding) selects the most probable token, while a topK of 3 selects from the top 3 probabilities. This helps narrow down the options for the next token during the generation process.

• max_output_tokens

  • Specifies the maximum number of tokens that can be generated in the response. A token is approximately four characters. 100 tokens correspond to roughly 60-80 words

Safety Settings

Safety Category	Description
Harassment	Negative or harmful comments targeting identity and/or protected attributes.
Hate Speech	Content that is rude, disrespectful, or profane.
Sexually explicit	Contains references to sexual acts or other lewd content.
Dangerous	Promotes, facilitates, or encourages harmful acts.

Threshold	Description
BLOCK_NONE	Always show regardless of probability of unsafe content
BLOCK_ONLY_HIGH	Block when high probability of unsafe content
BLOCK_MEDIUM_AND_ABOVE	Block when medium or high probability of unsafe content
BLOCK_LOW_AND_ABOVE	Block when low, medium or high probability of unsafe content
HARM_BLOCK_THRESHOLD_UNSPECIFIED	Threshold is unspecified, block using default threshold

Note: You have to give safety setting in the order of the json file, Do not change the given order of safety category. Just change the Threshold

Usage

1. Run the script directly

   go run main.go

   You should be in the same directory as the file main.go

2. Run the binary, Which is compiled

   ./gemini

3. If you are a Linux/Unix User you can move the gemini binary to /usr/local/bin/ and access gemini as a standard Command

   chmod +x gemini
   sudo mv gemini /usr/local/bin
   sudo mv Gemini /usr/local/etc  # Move the directory which has the config file to /usr/local/etc

   Now you can use the app as an ordinary terminal application

   gemini

Command-Line Arguments

• --text

  • Start Conversation with the Text-to-Text Mode. (Default)

• --image

  • Start Conversation with the Multi-Mode Model. This model will allow you to give inputs as Text as well as Image
    • When asking for the path of the image you can,
      • Type the path (absolute path)
      • Copy the image path to the Clipboard and hit Enter
      • Drag and drop the image to the Terminal

• -h, --help

  • Shows help menu and exit

Examples

go run main.go

---

go build -o gemini
./gemini

---

gemini

---

gemini --help

---

gemini --image

The same project in python can be found at: mr-alham/Google-Gemini-AI-on-the-Terminal, But it is deprecated.

License

This project is licensed under the MIT License.

Author

Written and maintaining by Alham

• GitHub : mr-alham
• Twitter (X) : @alham__aa
• Email : alham@duck.com