AI Agent Sandbox

A multi-agent system for Node.js programming assistance built with Google Agent Development Kit (ADK). This project features a coordinating agent (DevDuck) that manages two specialized sub-agents (Local Agent and Cerebras) for different programming tasks.

Architecture

The system consists of three main agents orchestrated by Docker Compose, which plays a primordial role in launching and coordinating all agent services:

🐙 Docker Compose Orchestration

• Central Role: Docker Compose serves as the foundation for the entire multi-agent system
• Service Orchestration: Manages the lifecycle of all three agents (DevDuck, Local Agent, and Cerebras)
• Configuration Management: Defines agent prompts, model configurations, and service dependencies directly in the compose file
• Network Coordination: Establishes secure inter-agent communication channels
• Environment Management: Handles API keys, model parameters, and runtime configurations

Agent Components

🦆 DevDuck (Main Agent)

• Role: Main development assistant and project coordinator
• Model: llama3
• Capabilities: Routes requests to appropriate sub-agents based on user needs

👨‍💻 Local Agent Agent

• Role: General development tasks and project coordination
• Model: mistral
• Specialization: Node.js programming expert for understanding code, explaining concepts, and generating code snippets

🧠 Cerebras Agent

• Role: Advanced computational tasks and complex problem-solving
• Model: Llama-4 Scout (llama-4-scout-17b-16e-instruct)
• Provider: Cerebras API
• Specialization: Node.js programming expert for complex problem-solving scenarios

Features

• Multi-agent coordination: Routing between specialized agents
• Node.js programming expertise: All agents specialize in Node.js development
• FastAPI web interface: RESTful API with web interface support
• Docker containerization: Easy deployment with Docker Compose
• Flexible model configuration: Support for multiple LLM providers (local and cloud)

Quick Start

Prerequisites

• [Docker Desktop] 4.43.0+ or [Docker Engine] installed.
• A laptop or workstation with a GPU (e.g., a MacBook) for running open models locally. If you don't have a GPU, you can alternatively use [Docker Offload].
• If you're using [Docker Engine] on Linux or [Docker Desktop] on Windows, ensure that the [Docker Model Runner requirements] are met (specifically that GPU support is enabled) and the necessary drivers are installed.
• If you're using Docker Engine on Linux, ensure you have [Docker Compose] 2.38.1 or later installed.

Configuration

1. You need a Cerebras API Key: https://cloud.cerebras.ai/
2. Create a .env file with the following content:

CEREBRAS_API_KEY=<your_cerebras_api_key>
CEREBRAS_BASE_URL=https://api.cerebras.ai/v1
CEREBRAS_CHAT_MODEL=llama-4-scout-17b-16e-instruct

look at the .env.sample file

✋ All the prompts are defined in the 🐙 compose file

Start the services

docker compose up
# if you updated the code, use --build

The application will be available at http://0.0.0.0:8000

---

📚 Dạy Học Online

Bên cạnh tài liệu miễn phí, mình còn mở các khóa học online:

• Lập trình web cơ bản → nâng cao
• Ứng dụng về AI và Automation
• Kỹ năng phỏng vấn & xây CV IT

Thông Tin Đăng Ký

• 🌐 Website: https://profile-forme.com
• 📧 Email: nguyentientai10@gmail.com
• 📞 Zalo/Hotline: 0798805741

---

💖 Donate Ủng Hộ

Nếu bạn thấy các source hữu ích và muốn mình tiếp tục phát triển nội dung miễn phí, hãy ủng hộ mình bằng cách donate.
Mình sẽ sử dụng kinh phí cho:

• 🌐 Server, domain, hosting
• 🛠️ Công cụ bản quyền (IDE, plugin…)
• 🎓 Học bổng, quà tặng cho cộng đồng

QR Code Ngân Hàng

Quét QR để ủng hộ nhanh:

QR Code ABBank

• Chủ tài khoản: Nguyễn Tiến Tài
• Ngân hàng: NGAN HANG TMCP AN BINH
• Số tài khoản: 1651002972052

---

📞 Liên Hệ

• 📚 Tiktok Dạy Học: @code.web.khng.kh
• 💻 GitHub: fdhhhdjd
• 📧 Email: nguyentientai10@gmail.com

Cảm ơn bạn đã quan tâm & chúc bạn học tập hiệu quả! Have a nice day <3!!