Echo

The Future of Desktop Computing

Echo is a multimodal AI agent that lives on your desktop.
It doesn't just chat—it acts. Using advanced speech-to-speech models and the Model Context Protocol (MCP), Echo listens to your voice, understands your intent, and controls your computer to get things done.

Features • Installation • Architecture • Documentation

---

What is Echo?

Echo bridges the gap between conversational AI and OS-level control. Most assistants are trapped in a browser tab. Echo integrates with your operating system, allowing it to:

• See your screen and understand context.
• Hear your voice with sub-second latency (Real-time API).
• Act on your apps, files, and workflows using specialized tools.

Whether you're managing a Google Classroom, automating a complex workflow, or building a presentation, Echo acts as your intelligent co-pilot.

Features

Native Voice Interaction

Speak naturally. Echo uses Gemini 2.0 Flash's native audio capabilities for fluid, interruptible, human-like conversation. No "wake words" or robotic pauses.

Full Desktop Control

Echo isn't limited to APIs. It can use your computer like a human:

• App Launching: "Open VS Code and Spotify."
• UI Interaction: Click, type, scroll, and navigate GUI applications.
• Screen Perception: It "sees" what you see to provide context-aware help.

Model Context Protocol (MCP)

Built on the open standard for AI tools. Echo connects to any MCP server:

• FileSystem: Read/Write files safely.
• Terminal: Execute commands and analyze output.
• Browser: Automate web research and tasks.
• Custom: Add your own tools easily.

Google Classroom Assistant

A dedicated module for educators:

• Course Management: Create courses, invite students, and manage rosters.
• Assignment Automation: Draft and publish assignments with attachments.
• Smart Forms: Generate quizzes and feedback forms automatically.

Transparent Reasoning

Watch Echo "think" in real-time. The UI visualizes the Chain of Thought (CoT), showing you exactly how the agent plans and executes complex tasks step-by-step.

---

Architecture

Echo uses a hybrid architecture to combine the best of web technologies and native performance.

graph TD
    User((User)) -->|Voice/Text| ElectronUI[Electron App / TUI]
    
    subgraph Frontend
        ElectronUI -->|WebSocket| Backend
        ElectronUI -->|Render| React[React UI]
    end
    
    subgraph Core ["Echo Backend (Python)"]
        Backend[FastAPI Server] -->|Orchestrate| Agent[Planner Agent]
        Backend -->|Stream| Voice[Gemini Live API]
        
        Agent -->|Think| CoT[Chain of Thought]
        Agent -->|Execute| Tools[Tool Manager]
    end
    
    subgraph Ecosystem ["MCP & APIs"]
        Tools <-->|Connect| MCP[MCP Servers]
        MCP -->|Control| Windows[Windows OS]
        MCP -->|Manage| Classroom[Google Classroom API]
        MCP -->|Access| Files[FileSystem]
    end

Loading

---

Quick Start

Prerequisites

• Python 3.10+
• Node.js 18+
• Google Gemini API Key (with Live API access)

Installation

1. Clone the repository

   git clone https://github.com/your-org/echo-desktop-agent.git
   cd echo-desktop-agent

2. Set up the environment Echo uses uv for fast Python package management (optional but recommended).

   # Install dependencies
   pip install -r requirements.txt
   
   # Or with uv
   uv sync

3. Configure Credentials Create a .env file in the root directory:

   GEMINI_API_KEY=your_api_key_here
   # Optional: For Classroom features
   GOOGLE_CLIENT_ID=...
   GOOGLE_CLIENT_SECRET=...

Running Echo

Option A: Electron App (Recommended)

The full visual experience with Voice UI and Chain-of-Thought visualization.

# Terminal 1: Start the backend
python src/backend/main.py

# Terminal 2: Start the UI
cd electron-app
npm install
npm start

Option B: Terminal Interface (TUI)

A lightweight, hacker-friendly interface for the terminal.

# Voice Mode (Interactive)
python TUI.py --mode voice

# Fast Command Mode
python TUI.py --command "Open Notepad and type Hello World"

---

Documentation

• Gemini Live MCP: Detailed guide for the web frontend and Google Classroom integration.
• Quickstart Guide: Extended setup instructions.
• MCP Configuration: Configure connected tool servers.

---

Project Structure

Echo/
├── TUI.py                 # Terminal User Interface entry point
├── electron-app/          # Desktop UI (Node.js/React)
├── gemini_live_mcp/       # Next.js Web Frontend & Classroom Module
├── src/
│   ├── agent/             # Core Agent Logic (Planner, Executor)
│   ├── tools/             # Native Tool Implementations
│   └── utils/             # Helpers for Audio, MCP, Logging
└── tests/                 # Unit and Integration Tests

---

Contributing

Echo is currently a private research project. Contributions are limited to the core team.

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

---

_{Built with by the Echo Team. Powered by Google DeepMind.}