Docker MCP Gateway
Docker’s official gateway for bundling and controlling container-isolated MCP servers for AI clients.
- Skill Road
- Docker MCP Gateway
Categories
Description
Docker MCP Gateway is Docker’s open-source, client-facing gateway implementation for the Model Context Protocol. It is neither just an SDK nor one domain-specific tool: an AI client connects to a gateway that starts registered MCP servers, aggregates their tools, routes requests, and controls credential access. The direct product link leads to Docker’s official documentation, while source and technical details are available in the linked GitHub repository. This makes the gateway particularly useful for teams using more than one MCP server without configuring every connection, container runtime, and secret separately in Claude Code, Codex, or Cursor.
A gateway instead of uncoordinated individual servers
Individual MCP servers are quick to install, but become difficult to manage at scale: every client configuration carries its own commands, tokens, images, and permissions. Docker MCP Gateway provides an intermediary layer. Docker says it manages server lifecycle, routing, credentials, and access control. A client commonly connects locally through stdio using docker mcp gateway run; the gateway then exposes the selected tools of the underlying servers. SSE and Streamable HTTP are also supported for networked cases. Docker documents operation through Docker Desktop with MCP Toolkit, as a docker mcp CLI plugin on Docker Engine, or through Docker Compose.
Containers, permissions, and secrets
The operational boundary is the practical value. According to the documentation, MCP servers run in isolated containers with restricted privileges, resource settings, and network controls. Host environment variables are not passed through by default. Secrets are scoped to one declaring server instead of being distributed to every container. --block-secrets is enabled by default and scans tool arguments and responses for secret-like values. This is a guardrail, not a promise that every secret will be detected: credentials should still be least-privilege and supplied only to the server that needs them.
For HTTP transports, the security documentation says the gateway requires a Bearer token by default, configured or generated through MCP_GATEWAY_AUTH_TOKEN. --allow-unauthenticated removes that boundary and is not a sensible default for reachable production endpoints. The health endpoint intentionally remains unauthenticated. Network egress is not globally blocked automatically either; users needing restrictions must deliberately configure blocking or allowlist options. Third-party images outside Docker’s signing namespaces do not receive the same signature verification as official images.
Clients and common uses
Docker explicitly documents integration with Claude Code, OpenAI Codex, and Cursor. A team can centrally activate a documentation server, database server, and cloud server while the agent still sees one MCP endpoint. That reduces duplicated configuration, but does not make permissions safer than the underlying servers and tokens allow. It is most useful for development teams standardizing agent workstations or operating several MCP servers through Docker Desktop and Compose.
License, version, and GitHub stars
The repository is MIT licensed. On 2026-09-07, GitHub’s API reported 1,556 stars; that point-in-time count is not evidence of security or quality. The newest tag was v0.43.3, while GitHub did not provide a separate latest-release record. The gateway itself is open source. Costs can arise from Docker products, infrastructure, or services used behind the gateway; their official provider terms apply.
Who benefits most
The gateway pays off when multiple MCP servers must be distributed repeatedly, containerized, and bounded clearly across coding clients. For one trusted local server, the additional layer is often unnecessary. Start with a small read-only toolset, inspect images and logs, then expand access only after a traceable test. That keeps the gateway an operational simplification rather than a new uncontrolled collection point for access.
Requirements
Docker Desktop with MCP Toolkit or Docker Engine/Compose plus an MCP client.
Installation instructions
Quick start
Enable Docker MCP Toolkit or run the gateway with Docker Compose. Connect the client through docker mcp gateway run and initially enable only approved read-only servers.
docker mcp gateway run
What the Docker MCP Gateway is and why it matters
The Docker MCP Gateway is, according to the vendor, Docker's open-source solution for centrally running and managing multiple Model Context Protocol (MCP) servers. MCP is an open protocol that lets AI applications such as Claude Desktop, Claude Code, Cursor, or VS Code access external tools, data sources, and APIs in a structured way. Without a gateway, every single client application has to know on its own how and where to start an MCP server, which credentials it needs, and how to secure it. The Gateway centralizes that job: according to Docker, it acts as a proxy between clients and servers, handling configuration, credentials, and access control, and automatically starting the actual MCP servers as isolated Docker containers when needed. This is particularly relevant for teams running several AI tools in parallel who don't want to maintain a separate, potentially insecure server configuration for each application. Technically, the project consists of the docker-mcp CLI plugin, which either runs embedded in Docker Desktop with the MCP Toolkit enabled, or standalone on Docker Engine without a desktop UI, for example under WSL2 or in containerized environments.
Step 1: Prerequisites
For the simplest path, Docker states that Docker Desktop version 4.59 or later with the MCP Toolkit enabled is required; in that case the Gateway runs automatically in the background and doesn't need to be started manually. Anyone wanting to run the Gateway without Docker Desktop, for instance on a server running Docker Engine, needs the CLI plugin installed separately. Building from source additionally requires Go version 1.24 or later. To operate it at all, a working Docker installation is sufficient, since the MCP servers run as separate containers with restricted privileges. It's also important to have a clear picture of which MCP servers from the Docker MCP Catalog are actually needed, because every additional server adds attack surface and resource consumption.
Step 2: Step-by-step setup
Docker Desktop users first enable the MCP Toolkit in settings and configure the desired servers from the catalog there; the Gateway then runs implicitly in the background. For standalone operation, the documentation describes downloading the current binary from the GitHub releases page and moving it to the OS-appropriate location for CLI plugins — on Linux and macOS, for example, to ~/.docker/cli-plugins/docker-mcp, and on Windows to the corresponding folder in the user profile. After setting execute permissions, the docker mcp --help command can be run to verify the installation. Servers are then organized into so-called profiles, which are reusable collections of MCP servers that can come from the Docker MCP Catalog, custom OCI images, the public MCP registry, or local files. A profile is created with the docker mcp profile create command and then connected to a client such as Claude Code or VS Code. In environments without Docker Desktop, such as plain Docker CE, the vendor notes that the profiles feature must first be explicitly enabled with docker mcp feature enable profiles.
Security and best practices
The Gateway's central security concept is isolation: each MCP server runs in its own container with restricted privileges, restricted network access, and limited resource usage, rather than running directly on the host system with the calling user's permissions. This reduces the risk that a compromised or buggy server can reach the rest of the system. According to the vendor, credentials such as API keys are handled through Docker Desktop's secrets management rather than stored as plaintext environment variables, which lowers the risk of accidental leaks into logs or version control. The Gateway also provides built-in logging and call tracing, so it's possible to review afterward which tool was called when and with what data. Anyone running the Gateway in production should still only allow servers from trusted sources in their profile and regularly review what permissions individual servers actually request, since MCP servers can in principle gain access to sensitive data and systems.
Practical example and limitations
A typical scenario is a development team that wants to reach GitHub, a database, and an internal search server simultaneously from several different AI clients. Instead of configuring each application separately, a shared profile containing all three servers is created and connected to every client, producing consistency and central control. It's worth noting that, according to Docker's own documentation, certain advanced features of the Gateway are currently part of what's called Docker AI Governance and are invite-only, while the core of the CLI plugin remains openly available. The Gateway also does not replace scrutiny of individual MCP servers' actual behavior: isolation protects against system-level access, but not against a server itself returning incorrect or unwanted results. Anyone using MCP servers in production should therefore still critically review the data returned by individual servers before feeding it into automated workflows.
docker mcp gateway run
Authentication
Stdio runs locally. HTTP requires a Bearer token by default.
Required access permissions
Permissions derive from connected servers, secrets, mounts, and network rules.
Transmitted or stored data
Tool calls are routed to connected servers. Docker documents metadata logging by default, but no blanket data-residency guarantee.
Security risks
Unreviewed images, broad tokens, mounts, or network access enlarge the attack surface. Never expose HTTP unintentionally without authentication.
License and costs
- License
- MIT
- Cost
- free
The gateway is MIT licensed. See the respective providers for infrastructure, Docker offering, and connected-service terms.
Alternatives
Not recorded yet.
At a glance
- Provider
- Docker
- Status
- Official server
- Deployment
- Local and remote
- Current version
- 0.43.3
- GitHub stars
- 1,637
- Last reviewed
- 07.09.2026
Repository and documentation
Categories
Supported clients
Related guides
Guides and background related to this entry.