docker-agent
AI Agent Builder and Runtime by Docker Engineering
What is docker-agent?
AI Agent Builder and Runtime by Docker Engineering
Overview of docker-agent
What it is
Docker Agent is an AI agent builder and runtime developed by Docker Engineering. It allows users to create, run, and share intelligent AI agents using a declarative YAML configuration. The tool supports multi-agent orchestration and a rich tool ecosystem, and it can be used as a Docker CLI plugin via the docker agent command.
Key features
- Multi-agent architecture — Create teams of specialized agents that delegate tasks automatically
- Rich tool ecosystem — Built-in tools plus any MCP server (local, remote, or Docker-based)
- AI provider agnostic — Supports OpenAI, Anthropic, Gemini, AWS Bedrock, Mistral, xAI, Docker Model Runner, and more
- YAML configuration — Declarative, versionable, and shareable agent definitions
- Advanced reasoning — Built-in think, todo, and memory tools
- RAG — Pluggable retrieval with BM25, embeddings, hybrid search, and reranking
- Package and share — Push agents to any OCI registry, pull and run them anywhere
Use cases
Docker Agent can be used to build AI assistants that collaborate on complex problems. It is suitable for creating specialized agent teams, integrating with various AI model providers, and deploying agents in different environments. The YAML-based configuration makes it easy to version and share agent definitions across teams or publicly via OCI registries.
Getting started
Installation options include Docker Desktop (version 4.63+), Homebrew (brew install docker-agent), or binary releases from GitHub. After installation, set at least one API key (e.g., export OPENAI_API_KEY=sk-...) or use Docker Model Runner for local models. Quick start commands:sh# Run the default agentdocker agent run# Run from an OCI registrydocker agent run myorg/agent:tag# Generate a new agent interactivelydocker agent new# Run your own configdocker agent run agent.yaml
License and pricing
Docker Agent is open source under the Apache-2.0 license. No pricing information is stated in the provided materials.
Based on the project's public README and repository metadata.