This repository provides a collection of Docker Compose configurations for building and running AI agents, simplifying the orchestration of LLMs, tools, and agent runtimes. It targets developers and researchers looking to experiment with and deploy multi-agent systems using readily available open-source components and Docker.

How It Works

The project leverages Docker Compose to define and manage multi-container applications for AI agents. Each demo is a self-contained project with a compose.yaml file, allowing users to easily spin up complex agent setups. It supports both local model execution via Docker Model Runner (DMR) and integration with cloud-based LLMs like OpenAI, offering flexibility in deployment and cost management.

Quick Start & Requirements

• Install: docker compose up --build within a demo directory.
• Prerequisites: Docker Desktop 4.43.0+ or Docker Engine with GPU support enabled and necessary drivers. Docker Compose 2.38.1+ for Linux.
• OpenAI Integration: Create a secret.openai-api-key file.
• Demos: A2A Multi-Agent Fact Checker, Agno GitHub Issue Summarizer, Vercel AI-SDK Chat-UI, CrewAI Marketing Strategy Agent, ADK Multi-Agent Fact Checker, ADK & Cerebras Golang Experts, LangGraph SQL Agent, Embabel Travel Agent, Spring AI Brave Search, ADK Sock Store Agent, Langchaingo DuckDuckGo Search.

Highlighted Details

• Supports local execution of various LLMs (Qwen, Llama, Gemma) via Docker Model Runner.
• Integrates with external tools and APIs like DuckDuckGo, GitHub, Brave Search, and email services.
• Demonstrates multi-agent architectures using frameworks like CrewAI and LangGraph.
• Offers examples for single-agent setups and complex multi-agent interactions.

Maintenance & Community

No specific contributors, sponsorships, or community links (Discord/Slack) are mentioned in the README.

Licensing & Compatibility

Dual-licensed under Apache License 2.0 or MIT License. Individual examples may have their own licenses that must be respected. Compatible with commercial use under the terms of either license.

Limitations & Caveats

The README does not detail specific limitations or known issues. GPU support is highly recommended for local model execution, and users must ensure their Docker environment meets the specified requirements.