Gemini Chat CLI

A command-line interface for chatting with Google's Gemini AI model, built in Rust.

Features

• Interactive Chat: Real-time conversation with Gemini AI
• Streaming Responses: See responses as they're generated (default mode)
• Non-streaming Mode: Get complete responses at once
• Built-in Commands: Help, clear screen, quit, toggle streaming
• Error Handling: Robust error handling for API issues
• Environment Configuration: Secure API key management

Setup

1. Clone and build the project:

   cargo build

2. Set up your API key:

   cp .env.example .env
   # Edit .env and add your Gemini API key

3. Get a Gemini API key:

   • Go to Google AI Studio
   • Create a new API key
   • Copy it to your .env file

Usage

Start the chat:

cargo run

Available Commands:

• /help or /h - Show help message
• /clear or /cls - Clear the screen
• /quit, /exit, or /q - Exit the chat
• /stream - Toggle between streaming and non-streaming mode

Example Session:

Gemini Chat CLI
===============
Type your message and press Enter to send.
Commands:
  /help    - Show this help
  /clear   - Clear the screen
  /quit    - Exit the chat
  /stream  - Toggle streaming mode (default: on)

You: Hello! Can you explain what Rust is?
Gemini: Rust is a systems programming language that focuses on safety, speed, and concurrency...

You: /stream
Streaming mode: OFF

You: What are the main benefits of Rust?
Gemini: The main benefits of Rust include memory safety, performance, and fearless concurrency...

You: /quit
Goodbye!

Architecture

Core Components:

1. GeminiClient (src/gemini.rs):

   • Handles API communication with Google's Gemini service
   • Supports both streaming and non-streaming responses
   • Implements proper error handling and response parsing

2. Main CLI (src/main.rs):

   • Interactive command-line interface
   • User input handling and command processing
   • Response display and formatting

3. Additional Tools:

   • src/bin/rag.rs - RAG (Retrieval-Augmented Generation) implementation
   • src/bin/sse_client.rs - Server-Sent Events client for testing

Key Features:

• Async/Await: Built on Tokio for efficient async operations
• Streaming Support: Real-time response streaming using Server-Sent Events
• Error Handling: Comprehensive error handling with anyhow
• Environment Config: Secure API key management with dotenv

Dependencies

• tokio - Async runtime
• reqwest - HTTP client for API calls
• serde - JSON serialization/deserialization
• anyhow - Error handling
• futures - Stream processing
• dotenv - Environment variable management
• crossterm - Terminal manipulation

Configuration

The application uses environment variables for configuration:

GEMINI_API_KEY=your-actual-api-key-here

Development

Build:

cargo build

Run:

cargo run

Run specific binaries:

# RAG demo
cargo run --bin rag

# SSE client test
cargo run --bin sse_client

API Integration

The chat CLI integrates with Google's Gemini API using:

• Model: gemini-2.0-flash-exp (configurable)
• Endpoint: https://generativelanguage.googleapis.com/v1beta/models
• Features: Streaming and non-streaming content generation

Request Configuration:

• Temperature: 0.7
• Top-P: 0.95
• Top-K: 40
• Max Output Tokens: 2048

Troubleshooting

Common Issues:

1. API Key Error:

   GEMINI_API_KEY environment variable not set
   

   Solution: Copy .env.example to .env and add your API key.

2. Build Errors: Make sure you have Rust installed and run cargo build.

3. Network Issues: Check your internet connection and API key validity.

License

This project is open source. Feel free to use and modify as needed.