Skip to main content

deploy

Take an app live from chat: publish_app uploads the app's production build to its own <name>.digitorn.app URL, deployment_status reports whether it is live and its URL. The agent never leaves the conversation to hit a button.

Exact tools and parameters

catalog.describe_module deploy returns the real tools and every parameter, from the live registry.

Tools, at a glance: publish_app, deployment_status.

Gotchas that matter​

  • Build first. publish_app only uploads an existing production build - it never builds for you. Run the build (so a dist/ exists) with bash before publishing.
  • Redeploys overwrite in place and keep the same URL.
  • On the free plan a user may publish 2 apps; a 3rd returns an upgrade prompt - relay it in one sentence, don't retry.
  • Connecting a personal domain stays a guided UI step (only the domain owner can add DNS records); it is not a tool.

A working app​

Compiles as-is. The brain uses a Digitorn gateway model (the default - no API key of your own needed); see the note below to use your own key or a local model instead.

yaml
app:
  app_id: publisher
  name: Publisher
runtime:
  entry_agent: main
agents:
  - id: main
    system_prompt: |
      Build the app for production with bash, then publish it to a live URL.
      Report the URL and its status. If publishing returns an upgrade prompt,
      relay it plainly instead of retrying.
    brain:
      provider: openai
      backend: openai_compat
      model: mimo-v2.5-free   # a model the Digitorn gateway serves
      config:
        api_key: placeholder  # ignored in gateway mode
tools:
  modules:
    deploy: {}
    bash: {}
  capabilities:
    grant:
      - module: bash
        tools: [run]
      - module: deploy
        tools: [publish_app, deployment_status]
  • bash - runs the production build publish_app uploads