Skip to main content
GET
/
v1
/
agentic
/
documents
/
tables
/
{model_id}
Get Table
curl --request GET \
  --url https://api.example.com/v1/agentic/documents/tables/{model_id}
{
  "dataset_id": "3c90c3cc-0d44-4b50-8888-8dd25736052a",
  "schema": {
    "skill": "<string>",
    "use_case": "<string>",
    "questions": [
      {
        "id": "<string>",
        "question": "<string>",
        "rationale": "<string>",
        "unit": "<string>",
        "enum": [
          "<string>"
        ]
      }
    ],
    "metadata": {}
  },
  "document_ids": [
    "<string>"
  ],
  "failed_document_ids": [
    "<string>"
  ],
  "failures": [
    {
      "document_id": "<string>",
      "error": "<string>"
    }
  ]
}

Headers

​
authorization
string | null
​
refresh-token
string | null

Path Parameters

​
model_id
string<uuid>
required

Response

Successful Response

GET /v1/agentic/documents/tables/{dataset_id} -- the tabularized result.

The tabularized rows themselves live in object storage under the returned dataset_id (same prefix used by upload-sourced datasets), so this response is metadata-only: enough to construct a :class:outerproduct.agentic.documents.DocumentDataset handle without materializing rows on the client.

document_ids lists only the docs that produced a row. failed_document_ids lists docs that failed extraction (a row was not produced); failures carries the per-doc error message.

​
dataset_id
required
​
schema
Schema · object
required

A frozen list of :class:Question for one document type and one use case.

​
document_ids
string[]
required
​
failed_document_ids
string[]
​
failures
TabularizeFailure · object[]