Self configuring config

A Configorama file can be self configuring because Configorama can inspect it before resolving it, build a graph of the values it depends on, and use that graph to walk a human or agent through the setup process.

That graph includes environment variables, CLI options, file references, self references, defaults, type filters, allowed values, sensitivity, and prompt text. The setup wizard turns those dependencies into questions, then resolves the config with the answers for the current run.

Run it

configorama setup config.yml

Use npx when Configorama is not installed in the project:

npx --yes configorama setup config.yml

You can also run the same flow through the resolve command:

configorama config.yml --setup

Add prompt text

The wizard is strongest when the config explains its own dependencies. Use help() for short prompt text, or use comments when the prompt needs examples, source notes, grouping, or sensitivity.

service: billing-api
stage: ${option:stage | oneOf("dev", "staging", "prod") | help("Deployment stage")}
 
database:
  host: ${env:DB_HOST | help("Database host")}
  port: ${env:DB_PORT, 5432 | Number | help("Database port")}
 
secrets:
  # Stripe live secret key
  # @from Stripe dashboard > Developers > API keys
  # @example sk_live_...
  # @sensitive true
  stripeKey: ${env:STRIPE_SECRET_KEY}

In this example, stage becomes a select prompt because oneOf(...) defines the allowed values in the dependency graph. database.port is validated as a number. secrets.stripeKey is treated as sensitive because the annotation says so, and names such as secret, password, token, and key are also masked by default.

What it prompts for

The wizard groups dependency inputs by source:

Defaults are shown as placeholders and used when the prompt is left blank. Type filters such as Number, boolean filters, JSON/object filters, and array filters are used for validation and normalization, so the prompt behavior follows the same dependency data that resolution uses.

For CI and agents, use configorama inspect config.yml --view requirements instead. The requirements output is the durable JSON contract behind the wizard, and it is easier to review, diff, and store as an artifact.

See inspect required inputs for the JSON shape and filters and functions for help() and oneOf(...).