Repolode
Submit
docker

docker/docker-agent

A Go-based CLI plugin for Docker that builds, runs and shares AI agents defined in declarative YAML, with MCP tools and multi-agent delegation.

View on GitHubWebsite
Stars
4,662
Forks
553
Open issues
72
Last commit
2 days ago

What is it?

Docker Agent is a runtime for building AI agents without writing code. Agents are described in a YAML file that specifies the model, instructions, and toolsets, and can be run locally with `docker agent run`. It runs as a docker CLI plugin, so the same command works after installation through Docker Desktop, Homebrew, or a binary release.

Agents can be organized into teams of specialized sub-agents that delegate tasks to one another. Tools come from built-in capabilities such as think, todo, and memory, plus any MCP server, whether local, remote, or Docker-based. A RAG component supports BM25, embeddings, hybrid search, and reranking.

The project works with multiple model providers, including OpenAI, Anthropic, Gemini, AWS Bedrock, Mistral, xAI, and Docker Model Runner for local models. Agent configurations can be pushed to and pulled from any OCI registry.

Key features

  • 01Declarative YAML agent definitions with model, instructions and toolsets
  • 02Multi-agent teams that delegate tasks to specialized sub-agents
  • 03Built-in think, todo and memory tools plus any MCP server
  • 04RAG with BM25, embeddings, hybrid search and reranking
  • 05Support for many model providers, including local models via Docker Model Runner
  • 06Push and pull agent packages to OCI registries

Use cases

  • Building a research assistant that uses web search through an MCP server
  • Creating a team of agents where one coordinates and others handle specialized subtasks
  • Sharing a ready-made agent configuration with teammates via an OCI registry
  • Running agents against local models for private workloads

Quick start

  1. 1

    Install with Homebrew

    brew install docker-agent
  2. 2

    Set at least one model provider API key

    export OPENAI_API_KEY=sk-...
  3. 3

    Run the default agent

    docker agent run

Things to know

  • Most setups need an API key from a cloud model provider, or a local model via Docker Model Runner.
  • The tool collects anonymous usage data by default, as described in its telemetry documentation.

In collections

Multi-agent orchestration frameworksOpen-source AI agent frameworks & appsMCP servers

Topics

#agents#ai-agents#llm#mcp#yaml#multi-agent#docker#rag#ai

FAQ

What is docker-agent?
A Go-based CLI plugin for Docker that builds, runs and shares AI agents defined in declarative YAML, with MCP tools and multi-agent delegation. Docker Agent is a runtime for building AI agents without writing code. Agents are described in a YAML file that specifies the model, instructions, and toolsets, and can be run locally with `docker agent run`. It runs as a docker CLI plugin, so the same command works after installation through Docker Desktop, Homebrew, or a binary release.
Is docker-agent free to use? What license does it have?
docker-agent is open source under the Apache-2.0 license. Check the license terms before using it in commercial products.
Is docker-agent still maintained?
Actively maintained. The last commit to docker-agent was on Nov 5, 2026.
What can I use docker-agent for?
Building a research assistant that uses web search through an MCP server Creating a team of agents where one coordinates and others handle specialized subtasks Sharing a ready-made agent configuration with teammates via an OCI registry Running agents against local models for private workloads
How do I install docker-agent?
Install with Homebrew: brew install docker-agent Set at least one model provider API key: export OPENAI_API_KEY=sk-... Run the default agent: docker agent run
Where can I get docker-agent?
The source code and installation instructions are on GitHub: https://github.com/docker/docker-agent

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects