Trace every answer back to the source

Every field ADE extracts comes with its exact page and bounding box — no black-box outputs.

This module

Use visual grounding to verify extractions. trace any extracted field back to its exact location on the page.

What is visual grounding?

Every chunk and extracted field in an ADE response includes grounding: the page number and the bounding box where that content sits in the original document. Grounding turns ADE output from a black-box JSON blob into a verifiable, auditable answer — you can always point at exactly where the value came from.

Why it matters

✓Verification & trust. Reviewers can jump from an extracted value straight to the page region it came from.
✓Audit trails for compliance. Grounding gives you a provable link between extracted data and the source document.
✓Faster debugging. When extraction is wrong, the bounding box tells you whether it's an OCR, layout, or schema problem.
✓Better review UIs. Build click-to-highlight, redaction, and side-by-side review experiences out of the box.

What you get in the response

Every chunk carries its page and normalized bounding box, and the response also includes a top-level grounding dictionary with richer details (type, confidence) keyed by element ID.

PER-CHUNK GROUNDING

{
  "id": "7d58c5cf-e4f5-4a7e-ba34-0cd7bc6a6506",
  "type": "text",
  "markdown": "Invoice Total: $4,820.00",
  "grounding": {
    "page": 1,
    "box": { "left": 0.612, "top": 0.843, "right": 0.891, "bottom": 0.872 }
  }
}

TOP-LEVEL GROUNDING — KEYED BY ELEMENT ID

"grounding": {
  "7d58c5cf-...": {
    "type": "chunkText", "page": 1, "confidence": 0.95,
    "box": { left: 0.61, top: 0.84, right: 0.89, bottom: 0.87 }
  },
  "0-1": { "type": "table", "page": 0, "box": { ... } },
  "0-2": { "type": "tableCell", "position": { row: 0, col: 0, ... } }
}