PB_MARKDOWN_HTML_LEAKAGE

What happened

The Markdown input contained raw HTML tags that were passed through to the renderer. Paperbase allows raw HTML in Markdown by default (which is usually desired — for embedding charts, custom layouts, etc.). This warning fires when the HTML appears unintentional, such as when LLM-generated Markdown accidentally includes literal <div> tags meant to be code samples.

How to fix

If the HTML is intentional, ignore this warning — raw HTML is a supported feature.

If the HTML is unintentional (e.g., LLM output that escaped code fences):

Option 1: Wrap HTML in code fences so it renders as a code block:

Here is a div:
 
```html
<div>content</div>

**Option 2:** Enable HTML sanitization:
```ts
const result = await paperbase.pdf.generate({
  input: { type: "markdown", content: md },
  template: "report",
  options: { sanitize_html: true }, // strips raw HTML from Markdown
});

Notes

sanitize_html: true strips all raw HTML from Markdown input. Use this when your content comes from untrusted sources or when you want to guarantee that only Markdown features are used.