SSHAI - SSH AI Assistant

English | 简体中文

An intelligent AI assistant program that provides AI model services through SSH connections, allowing you to use AI assistants in any SSH-supported environment.

✨ Key Features

• 🔐 Secure SSH Connection - Encrypted AI service access through SSH protocol
• 🔑 Flexible Authentication - Support for password authentication, SSH key-based passwordless login, and passwordless mode
• 🗝️ SSH Keys Support - Support for multiple SSH public keys for passwordless login, compatible with RSA, Ed25519, and other key types
• 🤖 Multi-Model Support - Support for DeepSeek, Hunyuan, and other AI models
• 💭 Real-time Thinking Display - Real-time display of thinking processes for models like DeepSeek R1
• 🛠️ MCP Tool Support - Support for Model Context Protocol, integrating various external tools and services
• 🎨 Beautiful Interface - Colorful output, animations, and ASCII art
• ⚙️ Flexible Configuration - Support for dynamic configuration file specification (-c parameter) and complete YAML configuration
• 🌐 Multi-language Support - Support for Chinese and English interfaces
• 📝 Custom Prompts - Configurable AI prompt system
• 🚀 Startup Welcome Banner - Beautiful welcome banner displayed on program startup
• 🏗️ Modular Design - Clean code architecture, easy to extend

🚀 Quick Start

1. Download and Build

# Clone the project
git clone https://github.com/sshllm/sshai.git
cd sshai

# Build the program
make build
# or
go build -o sshai cmd/main.go

2. Configuration Setup

Edit the config.yaml file and set your API key:

# API Configuration
api:
  base_url: "https://api.deepseek.com/v1"
  api_key: "your-api-key-here"
  default_model: "deepseek-v3"

# Server Configuration
server:
  port: 2213
  welcome_message: "Welcome to SSHAI!"

# Authentication Configuration (Optional)
auth:
  password: ""  # Empty = no password authentication
  login_prompt: "Please enter password: "
  # SSH public key passwordless login configuration (only effective when password is set)
  authorized_keys:
    - "ssh-rsa AAAAB3NzaC1yc2EAAAADAQABAAABAQC... user@hostname"
    - "ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAI... user2@hostname"
  authorized_keys_file: "~/.ssh/authorized_keys"  # Optional: read public keys from file

# Custom Prompt Configuration
prompt:
  system_prompt: "You are a professional AI assistant, please answer questions in English."
  stdin_prompt: "Please analyze the following content and provide relevant help or suggestions:"
  exec_prompt: "Please answer the following question or execute the following task:"

# MCP Tool Configuration
mcp:
  enabled: true  # Enable MCP functionality
  refresh_interval: 300  # Tool list refresh interval (seconds)
  servers:
    # Stable services using uvx (recommended)
    - name: "time"
      transport: "stdio"
      command: ["uvx", "mcp-server-time"]
      enabled: true
    - name: "fetch"
      transport: "stdio"
      command: ["uvx", "mcp-server-fetch"]
      enabled: true
    # Services using npx (may require longer startup time)
    - name: "bing"
      transport: "stdio"
      command: ["npx", "bing-cn-mcp"]
      enabled: true

3. Run the Server

# Run directly (using default config.yaml)
./sshai

# Specify configuration file
./sshai -c config.yaml
./sshai -c /path/to/your/config.yaml

# Run in background
./sshai > server.log 2>&1 &

# Run with script
./scripts/run.sh

Command Line Parameters

• -c <config_file> - Specify configuration file path
  • If not specified, defaults to config.yaml in current directory
  • If configuration file doesn't exist, program will show error message and exit

# Usage examples
./sshai -c config.yaml              # Use config file in current directory
./sshai -c /etc/sshai/config.yaml   # Use config file with absolute path
./sshai                             # Default to config.yaml

4. Connect and Use

# Interactive mode
ssh user@localhost -p 2213

# Direct command execution
ssh user@localhost -p 2213 "Hello, please introduce yourself"

# Pipe input analysis
cat file.txt | ssh user@localhost -p 2213
echo "Analyze this code" | ssh user@localhost -p 2213

📁 Project Structure

sshai/
├── README.md              # Chinese documentation
├── README_EN.md           # English documentation
├── LICENSE                # Open source license
├── config.yaml           # Main configuration file
├── config-en.yaml        # English configuration file
├── go.mod                # Go module dependencies
├── Makefile              # Build script
├── cmd/                  # Program entry
│   └── main.go           # Main program file
├── pkg/                  # Core modules
│   ├── config/           # Configuration management
│   ├── models/           # Data models
│   ├── ai/               # AI assistant functionality
│   ├── ssh/              # SSH server
│   └── utils/            # Utility functions
├── docs/                 # Project documentation
├── scripts/              # Test and run scripts
└── keys/                 # SSH key files

🔧 Configuration Guide

API Configuration

Support for multiple API endpoint configurations:

api:
  base_url: "https://api.deepseek.com/v1"
  api_key: "your-deepseek-key"
  default_model: "deepseek-v3"
  timeout: 600

# Multiple API configuration
api_endpoints:
  - name: "deepseek"
    base_url: "https://api.deepseek.com/v1"
    api_key: "your-key"
    default_model: "deepseek-v3"
  - name: "local"
    base_url: "http://localhost:11434/v1"
    api_key: "ollama"
    default_model: "gemma2:27b"

Authentication Configuration

Password Authentication

auth:
  password: "your-secure-password"  # Set access password
  login_prompt: "Please enter password: "

SSH Public Key Passwordless Login

auth:
  password: "your-secure-password"  # Password must be set to enable SSH public key authentication
  login_prompt: "Please enter password: "
  # Method 1: Configure public key list directly
  authorized_keys:
    - "ssh-rsa AAAAB3NzaC1yc2EAAAADAQABAAABAQC... user@hostname"
    - "ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAI... user2@hostname"
  # Method 2: Read public keys from file
  authorized_keys_file: "~/.ssh/authorized_keys"

SSH Public Key Usage:

# Generate SSH key pair
ssh-keygen -t ed25519 -f ~/.ssh/sshai_key

# Connect using private key (passwordless login)
ssh -i ~/.ssh/sshai_key -p 2213 user@localhost

# View public key content (for configuration)
cat ~/.ssh/sshai_key.pub

Note:

• SSH public key authentication is only enabled when a password is set, providing additional security
• Supports multiple public keys simultaneously, compatible with RSA, Ed25519, ECDSA, and other key types
• After successful login, the program will automatically display built-in welcome information, no need to configure in the config file

Prompt Configuration

prompt:
  system_prompt: "You are a professional AI assistant..."
  stdin_prompt: "Please analyze the following content:"
  exec_prompt: "Please answer the following question:"

MCP Tool Configuration

SSHAI supports Model Context Protocol (MCP) for integrating various external tools and services:

mcp:
  enabled: true
  refresh_interval: 300  # Tool list refresh interval (seconds)
  servers:
    # Recommended using uvx (more stable, faster startup)
    - name: "time"
      transport: "stdio"
      command: ["uvx", "mcp-server-time"]
      enabled: true
    - name: "fetch"
      transport: "stdio"
      command: ["uvx", "mcp-server-fetch"]
      enabled: true
    - name: "filesystem"
      transport: "stdio"
      command: ["uvx", "mcp-server-filesystem", "/tmp"]
      enabled: true
    
    # npx services (may require longer startup time)
    - name: "bing"
      transport: "stdio"
      command: ["npx", "bing-cn-mcp"]
      enabled: true

MCP Tool Features:

• 🔧 Automatic Tool Discovery - Automatically detect and load tools provided by MCP servers
• 🔄 Smart Retry Mechanism - Connection optimization for package managers like npx
• ⏱️ Timeout Protection - Prevent tool calls from hanging, with connection timeout and retry support
• 🎯 Selective Output - Hide tool call processes in pipeline and command modes, showing only results
• 📊 Real-time Status Monitoring - Display tool connection status and execution progress

Supported Package Managers:

• uvx - Python package manager (recommended, more stable)
• npx - Node.js package manager (supported but may be slower)
• Direct commands - Pre-installed MCP servers

If you encounter npx server connection issues, please refer to the MCP NPX Troubleshooting Guide.

🧪 Testing

The project includes comprehensive test scripts:

# Basic functionality test
./scripts/test.sh

# SSH execution functionality test
./scripts/test_ssh_exec_final.sh

# Standard input functionality test
./scripts/test_stdin_feature.sh

# Authentication functionality test
./scripts/test_auth.sh

# DeepSeek R1 thinking mode test
./scripts/test_deepseek_r1.sh

# SSH Keys passwordless login functionality test
./scripts/test_ssh_keys.sh

📚 Documentation

• Configuration Guide - Detailed configuration instructions
• Usage Guide - Feature introduction and usage methods
• Architecture Documentation - Modular architecture design
• Authentication Configuration - SSH authentication configuration examples
• SSH Keys Guide - SSH public key passwordless login configuration guide
• MCP Usage Guide - Model Context Protocol tool integration guide
• MCP NPX Troubleshooting - NPX MCP server connection issue solutions

🤝 Contributing

Issues and Pull Requests are welcome!

1. Fork the project
2. Create a feature branch (git checkout -b feature/AmazingFeature)
3. Commit your changes (git commit -m 'Add some AmazingFeature')
4. Push to the branch (git push origin feature/AmazingFeature)
5. Open a Pull Request

📄 License

This project is licensed under the Apache License 2.0. See the LICENSE file for details.

🙏 Acknowledgments

Thanks to all developers and users who have contributed to this project!

---

Note: This project follows the Apache 2.0 open source license, welcoming both personal and commercial use.