Ten Agent

🚀 TEN Agent Component

TEN Agent is a real-time voice assistant platform integrating various AI technologies, supporting multimodal interaction and modular expansion.

🚀 Quick Start

✨ Features

• Real-time Voice Processing: Achieve voice interaction based on Deepgram's ASR and FishAudio's TTS.
• AI Inference Capability: Integrate the OpenAI API for intelligent conversations.
• Real-time Audio and Video: Utilize Agora.io to build real-time communication capabilities.
• Extension Plugin Support: Modular design for easy function expansion.

Project Architecture

📦 Installation

System Requirements

• Processor: Dual-core or higher
• Memory: 4GB or more
• Required Tools:
  • Docker container runtime environment
  • Node.js LTS version 18.x

Notes for macOS Configuration

For Macs with Apple chips, uncheck "Use Rosetta for x86/AMD64 emulation" in Docker settings. Note: This may slow down the build speed on ARM architectures, but the performance will be normal after deployment to an x64 server.

Installation Steps

1. Create a .env file

   cp ./.env.example ./.env
   

2. Configure Agora Keys Update the following variables in the .env file:

   AGORA_APP_ID=
   AGORA_APP_CERTIFICATE=
   

3. Start the Development Container

   docker compose up -d
   

4. Enter the Container Environment

   docker exec -it ten_agent_dev bash
   

5. Build the Agent Service

   task use
   

6. Run the Web Interface

   task run
   

7. Configure the Playground Open the address: localhost:3000 and select the required modules and plugins.

💻 Usage Examples

Basic Usage

Gemini Real-time Plugin

1. Open the Playground interface.
2. Select the "voice_assistant_realtime" diagram.
3. Select the Gemini real-time module.
4. Add the v2v extension and configure the API key.

📚 Documentation

Project Components

Join the Community

• Discord: Link
• GitHub Repository: Link

Contribution Guidelines

1. Code Style:
   • Follow the ES6 standard.
   • Use 2-space indentation.
2. Commit Guidelines:
   • Commit messages should include the type and purpose.
   • Provide relevant documentation support.

📄 License

This project is licensed under the MIT Open Source License.