Deployment

Deployment Options

dockerhub-api exposes its MCP server (console script dockerhub-mcp) four ways. Pick the row that matches where the server runs relative to your MCP client, then copy the matching mcp_config.json below. Replace the <your-…> placeholders with the values from the Configuration / Environment Variables section.

#	Option	Transport	Where it runs	mcp_config.json key
1	stdio	stdio	client launches a subprocess	command
2	Streamable-HTTP (local)	streamable-http	a local network port	command or url
3	Local container / uv	stdio or streamable-http	Docker / Podman / uv on this host	command or url
4	Remote URL	streamable-http	a remote host behind Caddy	url

1. stdio (local subprocess)

The client launches the server over stdio via uvx — best for local IDEs (Cursor, Claude Desktop, VS Code):

{
  "mcpServers": {
    "dockerhub-mcp": {
      "command": "uvx",
      "args": ["--from", "dockerhub-api", "dockerhub-mcp"],
      "env": {
        "DOCKERHUB_URL": "<your-dockerhub_url>",
        "DOCKER_REGISTRY_URL": "<your-docker_registry_url>",
        "DOCKER_SCOUT_URL": "<your-docker_scout_url>"
      }
    }
  }
}

2. Streamable-HTTP (local process)

Run the server as a long-lived HTTP process:

uvx --from dockerhub-api dockerhub-mcp --transport streamable-http --host 0.0.0.0 --port 8000
curl -s http://localhost:8000/health        # {"status":"OK"}

Then either let the client launch it:

{
  "mcpServers": {
    "dockerhub-mcp": {
      "command": "uvx",
      "args": ["--from", "dockerhub-api", "dockerhub-mcp", "--transport", "streamable-http", "--port", "8000"],
      "env": {
        "TRANSPORT": "streamable-http",
        "HOST": "0.0.0.0",
        "PORT": "8000",
        "DOCKERHUB_URL": "<your-dockerhub_url>",
        "DOCKER_REGISTRY_URL": "<your-docker_registry_url>",
        "DOCKER_SCOUT_URL": "<your-docker_scout_url>"
      }
    }
  }
}

…or connect to the already-running process by URL:

{
  "mcpServers": {
    "dockerhub-mcp": { "url": "http://localhost:8000/mcp" }
  }
}

3. Local container / uv

(a) Launch a container directly from mcp_config.json (stdio over the container — no ports to manage). Swap docker for podman for a daemonless runtime:

{
  "mcpServers": {
    "dockerhub-mcp": {
      "command": "docker",
      "args": [
        "run", "-i", "--rm",
        "-e", "TRANSPORT=stdio",
        "-e", "DOCKERHUB_URL=<your-dockerhub_url>",
        "-e", "DOCKER_REGISTRY_URL=<your-docker_registry_url>",
        "-e", "DOCKER_SCOUT_URL=<your-docker_scout_url>",
        "knucklessg1/dockerhub-api:latest"
      ]
    }
  }
}

(b) Run a local streamable-http container, then connect by URL:

docker run -d --name dockerhub-mcp -p 8000:8000 \
  -e TRANSPORT=streamable-http \
  -e PORT=8000 \
  -e DOCKERHUB_URL="<your-dockerhub_url>" \
  -e DOCKER_REGISTRY_URL="<your-docker_registry_url>" \
  -e DOCKER_SCOUT_URL="<your-docker_scout_url>" \
  knucklessg1/dockerhub-api:latest
# or, from a clone of this repo:
docker compose -f docker/mcp.compose.yml up -d

{
  "mcpServers": {
    "dockerhub-mcp": { "url": "http://localhost:8000/mcp" }
  }
}

(c) From a local checkout with uv:

uv run dockerhub-mcp --transport streamable-http --port 8000

4. Remote URL (deployed behind Caddy)

When the server is deployed remotely (e.g. as a Docker service) and published through Caddy on the internal *.arpa zone, connect with the "url" key — no local process or image required:

{
  "mcpServers": {
    "dockerhub-mcp": { "url": "http://dockerhub-mcp.arpa/mcp" }
  }
}

Caddy reverse-proxies http://dockerhub-mcp.arpa to the container's :8000 streamable-http listener; http://dockerhub-mcp.arpa/health returns {"status":"OK"} when the service is live.

MCP server (standalone)

pip install "dockerhub-api[mcp]"
export DOCKERHUB_USERNAME=youruser
export DOCKERHUB_TOKEN=dckr_pat_xxx
dockerhub-mcp --transport streamable-http --host 0.0.0.0 --port 8000

Transports: stdio (default), streamable-http, sse.

Docker Compose — MCP only

cp .env.example .env       # fill in DOCKERHUB_* values
docker compose -f docker/mcp.compose.yml up -d

Docker Compose — MCP + A2A agent

docker compose -f docker/agent.compose.yml up -d

This starts:

• dockerhub-api-mcp on port 8000 (streamable-http, /health endpoint)
• dockerhub-api-agent on port 9018 (A2A + AG-UI web interface), pointed at the MCP service via MCP_URL

MCP client registration

Register the server in any MCP-capable client using the template in mcp_config.json:

{
  "mcpServers": {
    "dockerhub-api": {
      "command": "uv",
      "args": ["run", "dockerhub-mcp"],
      "env": {
        "DOCKERHUB_USERNAME": "youruser",
        "DOCKERHUB_TOKEN": "dckr_pat_xxx",
        "DOCKERHUB_ALLOW_DESTRUCTIVE": "False"
      }
    }
  }
}

Security posture

• Leave DOCKERHUB_ALLOW_DESTRUCTIVE=False in shared deployments; enable it only for operator sessions that genuinely need deletes or org-settings writes.
• Prefer organization access tokens scoped with TYPE_REPO path globs over account-wide PATs for CI/CD.
• Eunomia policies and OTEL/Langfuse tracing are available through the agent-utilities server bootstrap (see .env.example).

Debug image

docker/debug.Dockerfile builds the package from the working tree with dev tooling (uv, ripgrep, starship) for interactive debugging:

docker build -f docker/debug.Dockerfile -t dockerhub-api:debug .
docker run --rm -it dockerhub-api:debug bash