docker-agent
Build, run, and share powerful AI agents with a declarative YAML or HCL config, rich tool ecosystem, and multi-agent orchestration — by Docker.
Declarative Agents
Define agents in simple YAML or HCL — their model, tools, behavior, and relationships. No framework code required.
Rich Tool Ecosystem
Built-in tools for filesystem, shell, memory, and more. Extend with any MCP server — local, remote, or Docker-based.
Multi-Agent Orchestration
Create teams of specialized agents that delegate tasks to each other automatically via a coordinator.
Multi-Model Support
Use OpenAI, Anthropic, Google Gemini, AWS Bedrock, Docker Model Runner, or bring your own provider.
Package & Share
Push agents to any OCI-compatible registry. Pull and run them anywhere — just like container images.
Multiple Interfaces
Beautiful TUI for interactive sessions, CLI for scripting, HTTP API for integrations, and MCP mode for interop.
See It in Action
Quick Example
Create a file called agent.yaml:
agents:
root:
model: anthropic/claude-sonnet-4-5
description: A helpful coding assistant
instruction: |
You are an expert developer. Help users write clean,
efficient code and follow best practices.
toolsets:
- type: filesystem
- type: shell
- type: think
Then run it:
# Launch the interactive TUI
docker agent run agent.yaml
# Or run a one-shot command
docker agent run --exec agent.yaml "Explain the code in main.go"
You can also write agents as .hcl files. See HCL Configuration for the block-based syntax.
Explore the Docs
Getting Started
Learn what docker agent is and get your first agent running in minutes.
Core Concepts
Understand agents, models, tools, and multi-agent orchestration.
Configuration
Complete reference for agent, model, and tool configuration.
Features
TUI, CLI, MCP mode, RAG, Skills, and distribution.
Model Providers
OpenAI, Anthropic, Gemini, AWS Bedrock, Docker Model Runner, and custom providers.
Community
Contributing guidelines, troubleshooting, and more.
