Skip to main content
The recommended way to run Superwire is the Docker executor. 
docker run --rm -p 13703:13703 rmilewski/superwire
The executor exposes:
  • POST /execute
  • POST /validate
  • POST /format
For /execute, set Accept: text/event-stream to receive Server-Sent Events.

​
Runtime requirements

Your executor host needs:
  • Docker or a compatible container runtime.
  • Network access to the model provider endpoint used by the workflow.
  • Network access to any MCP servers imported by the workflow.

​
Provider setup

A workflow creates provider instances supported by the executor. At the moment, executor provider support is OpenAI-compatible endpoints that implement /responses. 
secrets {
    api_key: string
}

provider llm from openai {
    endpoint: "https://api.openai.com/v1"
    api_key: secrets.api_key
}

model fast from llm {
    id: "gpt-4.1-mini"
}
The provider instance configures access to the backend. The model profile selects the provider-specific model ID used by agents.

​
Health checks

A deployment should verify that the executor is reachable before sending workflow traffic. The exact health endpoint depends on the executor build and deployment configuration.

​
See also