About Cuddler

Why we built Cuddler, how it works, and the schema model behind it.

Why Cuddler exists

Teams needed a reliable way to generate HTML from structured data without hidden formatting rules.
Cuddler separates that pipeline into two explicit schema contracts so producers and templates stay decoupled.

The Cuddler model

  • Data schema first: data.schema.json validates incoming JSON before rendering.
  • Template schema second: template.schema.json decides rendering only; it does not change data validation.
  • Shared versioning: data and template schemas move together for a document type.
  • Debuggable output: validation and render failures return structured diagnostics.

How it is used

  1. Build or collect JSON that matches the data schema (often via Codex skills and prompts).
  2. Validate with validateData(data, dataSchema).
  3. Render with renderDocument(data, templateSchema, options).
  4. Ship HTML and share failures quickly through trace/debug output.

Prompt and skill workflow

data.schema.prompt.txt is the source document used by a Codex skill.
Each new document type should keep prompt files, schema artifacts, and example JSON in the same folder under /schemas/*.

Ready to build with Cuddler? Request onboarding or send us a message.