Fresh

docker model start-runner

Back

Reference

• Get started
• Guides
• Manuals

docker model start-runner

Copy as Markdown

Open MarkdownAsk Docs AIClaudeOpen in Claude

| Description | Start Docker Model Runner (Docker Engine only) | | Usage | docker model start-runner |

Description

This command starts the Docker Model Runner without pulling container images. Use this command to start the runner when you already have the required images locally.

For the first-time setup or to ensure you have the latest images, use docker model install-runner instead.

Options

Option	Default	Description
--backend		Specify backend (llama.cpp|vllm|diffusers). Default: llama.cpp
--debug		Enable debug logging
--do-not-track		Do not track models usage in Docker Model Runner
--gpu	auto	Specify GPU support (none|auto|cuda|rocm|musa|cann)
--host	127.0.0.1	Host address to bind Docker Model Runner
--port		Docker container port for Docker Model Runner (default: 12434 for Docker Engine, 12435 for Cloud mode)
--proxy-cert		Path to a CA certificate file for proxy SSL inspection
--tls		Enable TLS/HTTPS for Docker Model Runner API
--tls-cert		Path to TLS certificate file (auto-generated if not provided)
--tls-key		Path to TLS private key file (auto-generated if not provided)
--tls-port		TLS port for Docker Model Runner (default: 12444 for Docker Engine, 12445 for Cloud mode)