Actions6
Document → Generate
AI-generatedSummary
Generate a PDF/A‑3 document from provided Markdown or HTML content.
Inputs
- documentFormat (required) — Source format; markdown is rendered via Markdig with raw HTML disabled; HTML is sanitized against ADR‑0013 allowlist.
- documentContent (required) — The document body; limited to 100KB; tenant-supplied HTML is sanitized — script/iframe/style/external img are stripped.
- documentTitle — Optional title used for rendered
, PDF metadata, and output filename. - documentAuthor — Optional author written into PDF metadata.
- documentWatermark — Optional CSS overlay for a watermark; multi‑page positioning may vary.
- documentTimestamp — Whether to apply an RFC 3161 trusted timestamp to the rendered PDF.
- documentEvidencePack — Whether to bundle a legal evidence pack ZIP (requires the evidence_packs feature on the tenant plan).
- documentAccessibilityLevel — PDF/UA‑1 produces a tagged, screen‑reader‑accessible PDF/A‑3u hybrid; required for EU Accessibility Act for B2C documents.
- documentIdempotencyKey — Optional Idempotency-Key header; if omitted the API hashes the payload for automatic deduplication.
Output shape
A single record containing the generated job identifier and status; typically includes 'jobId' and 'status' fields with a download URL once completed.
The operation returns synchronously only after the PDF generation finishes (usually 2‑6 seconds). The response body is the API payload; the actual binary PDF is not returned directly.
Examples
Example 1: Create a PDF/A‑3 from markdown content with title and author.
documentFormat: markdown, documentContent: '# Invoice', documentTitle: 'Invoice 2026-001', documentAuthor: 'Acme Corp', documentTimestamp: true, documentEvidencePack: false, documentAccessibilityLevel: 'None'
Example 2: Render HTML with a watermark and accessibility level.
documentFormat: html, documentContent: '