Eliza

A computer-use agent you control via Telegram. Give it tasks, approve commands, and let it work on your machine.

Features

• Computer Control: Execute bash commands, take screenshots, manage files — all from Telegram
• Visual Understanding: Bot can view and analyze screenshots to understand what's on your screen
• Multi-LLM Support: Powered by Google Gemini, OpenAI, Groq, or DeepSeek
• Safe Execution: All bash commands require your approval via inline buttons
• MCP Integration: Extend capabilities with Model Context Protocol servers
• Persistent Memory: Remembers conversation context across sessions

Built-in Tools

Tool	Description
bash	Execute shell commands with user approval via inline buttons
web_search	Search the web using DuckDuckGo
send_file	Send files/images to the user via Telegram
view_image	Load and analyze images from disk (screenshots, etc.)

Installation

Requires Python 3.13+ and uv.

# Clone the repository
git clone https://github.com/bnovik0v/eliza.git
cd eliza

# Install dependencies
uv sync

# Copy and configure environment
cp .env.example .env
# Edit .env with your API keys

Configuration

Create a .env file with the following variables:

# Required
TELEGRAM_TOKEN=your_telegram_bot_token
OWNER_USER_ID=your_telegram_user_id

# LLM Provider (choose one or more)
GOOGLE_API_KEY=your_google_api_key
OPENAI_API_KEY=your_openai_api_key
GROQ_API_KEY=your_groq_api_key
DEEPSEEK_API_KEY=your_deepseek_api_key

# Optional
DEFAULT_PROVIDER=google  # google, openai, groq, deepseek
GOOGLE_MODEL=gemini-3-flash-preview

Usage

# Run the bot
uv run eliza

Telegram Commands

• /start - Show bot info and available tools
• /clear - Clear conversation history
• /provider <name> - Switch LLM provider (google, openai, groq, deepseek)
• /tools - List available MCP tools

Example Tasks

See your screen:

"Take a screenshot and tell me what's happening"

Web search:

"What's the current Bitcoin price?"

System tasks:

"Check disk usage and clean up temp files"

Development:

"Run the tests and fix any failures"

File management:

"Find large files in my Downloads folder"

MCP Servers

Configure MCP servers in mcp_servers.json:

{
  "servers": {
    "filesystem": {
      "command": "npx",
      "args": ["-y", "@anthropic/mcp-filesystem", "/home/user"]
    }
  }
}

Project Structure

src/eliza/
├── bot/
│   ├── telegram_bot.py  # Main bot implementation
│   ├── bash_tool.py     # Bash execution with approval
│   └── security.py      # Owner-only decorator
├── llm/
│   ├── base.py          # Abstract LLM provider
│   ├── google.py        # Google Gemini provider
│   ├── openai.py        # OpenAI provider
│   ├── groq.py          # Groq provider
│   ├── deepseek.py      # DeepSeek provider
│   └── factory.py       # Provider factory
├── mcp/
│   ├── client.py        # MCP client manager
│   └── config.py        # MCP configuration
├── memory/
│   ├── base.py          # Memory interface
│   └── sqlite.py        # SQLite implementation
├── config.py            # Settings management
└── main.py              # Entry point

License

MIT