Skip to content

Output Formats

llms.txt lists every page for an agent

Each lexicon has its own output format, such as a CloudFormation template or Kubernetes YAML. The core evaluation pipeline produces a provider-agnostic resource graph, and the lexicon’s serializer converts it to that format.

  1. chant resolves all TypeScript expressions, imports and references into concrete values.
  2. Property names are converted from TypeScript conventions to the target format’s.
  3. The lexicon serializer writes the output.

Every lexicon serializer:

  • writes resources in a stable order, so the output is safe for version control and diff review
  • omits properties set to null from output objects, and keeps null values inside arrays
  • converts cross-resource references to the target format’s reference syntax

See your lexicon’s documentation for the specific output format.

chant