Document Detail

Get Document Detail

GET/api/v1/deployments/{deploymentId}/documents/{fileId}

Returns detailed information about a document including its metadata, row references, file conversions, and validation results.

Parameters

Name	Type	In	Required	Description
`deploymentId`	string	path	Yes	The model or deployment ID
`fileId`	string	path	Yes	The file ID

Request

const res = await fetch(
"https://app.limai.io/api/v1/deployments/dep_abc123/documents/file_001",
{
  headers: {
    Authorization: "Bearer YOUR_API_TOKEN",
  },
}
);
const doc = await res.json();

Response

Response200

{
  "id": "file_001",
  "name": "invoice.pdf",
  "originalExtension": "pdf",
  "status": "EXTRACTED",
  "extractionStatus": "COMPLETED",
  "validationStatus": "failed",
  "validationExpected": true,
  "validationAttemptId": "clx8f2k9a0001",
  "isReviewed": false,
  "createdAt": "2025-01-15T10:30:00.000Z",
  "updatedAt": "2025-01-15T10:31:00.000Z",
  "confidenceScore": 0.92,
  "validationFailCount": 1,
  "inExamplePool": false,
  "alias": null,
  "metadata": null,
  "conversions": [
    {
      "extension": "md",
      "status": "COMPLETED"
    }
  ],
  "rows": [
    {
      "rowId": "row_001",
      "status": "PENDING",
      "tableId": "tbl_001"
    }
  ],
  "validations": {
    "crossField": [
      {
        "ruleId": "vrule_9k1",
        "ruleName": "Net Amount vs Gross Amount",
        "passed": false,
        "fieldA": {
          "columnId": "col_net",
          "value": "1180.00"
        },
        "fieldB": {
          "columnId": "col_gross",
          "value": "1250.00"
        },
        "operator": "EQUALS",
        "expected": "1250.00",
        "actual": "1180.00"
      }
    ],
    "row": [],
    "document": []
  },
  "modelId": "dep_abc123",
  "viaDeploymentId": "dpl_prod01"
}

Response Fields

Field	Type	Description
`id`	string	The file ID.
`name`	string	Original file name.
`originalExtension`	string	Extension of the uploaded file.
`status`	string	Upload-level file status (`UPLOADED`, `EXTRACTED`, `FAILED`, …).
`extractionStatus`	`COMPLETED` \| `PROCESSING` \| `FAILED`	Derived from the latest extraction job for this document. A queued or in-flight job reports `PROCESSING`; a failed or stalled job reports `FAILED`; `COMPLETED` means the extraction finished (including documents that were never queued). Matches the `status` returned by Document Data.
`validationStatus`	`pending` \| `passed` \| `failed` \| `error` \| `null`	Document-wide aggregate over the model's validation rules and every custom validation check recorded for the document, across all attempts — it is not scoped to `validationAttemptId`. `failed` when a rule failed or a check reported findings, `pending` while a check is queued or running or a just-completed extraction job is still evaluating the model's rules, `error` when rule evaluation or a check could not run, `null` when the document has no rule verdict and no custom validations. Same value and precedence as `validations.status` on Document Data; see Validations.
`validationExpected`	boolean	Whether validation is configured for this document — validation rules on its model, enabled validation scripts, or a validator agent routed to the deployment. When `true` and `validationStatus` is `null`, nothing is recorded for the document yet: custom validation has not been dispatched, and the document holds no rule verdict — the state of a document extracted before its model had any rule.
`validationAttemptId`	string \| null	Identifier of the most recent validation attempt. Correlates this document with the `DOCUMENT_VALIDATION_STARTED`, `DOCUMENT_VALIDATION_FAILED`, and `DOCUMENT_VALIDATED` webhook events of the same attempt. `null` when the document's only validations are rules, which run synchronously and start no attempt. Use it for event correlation, not to interpret `validationStatus`: after a partial revalidation the two lanes can sit on different attempts, and `validationStatus` still answers the document-wide question "is this document validated?".
`isReviewed`	boolean	Whether a human has reviewed the extraction.
`createdAt`	string	ISO 8601 creation timestamp.
`updatedAt`	string	ISO 8601 last-update timestamp.
`confidenceScore`	number \| null	Overall extraction confidence, `null` when not computed.
`validationFailCount`	number	Number of failing checks recorded on the document, failing validation rules included.
`inExamplePool`	boolean	Whether the document is part of the model's example pool.
`alias`	string \| null	User-assigned alias.
`metadata`	object \| null	Metadata attached at upload time.
`conversions`	array	Derived artifacts, each `{extension, status}`.
`rows`	array	Extracted rows, each `{rowId, status, tableId}`.
`validations`	object	Validation rule results grouped by `crossField`, `row`, and `document`. See Rule results.
`modelId`	string	The model this request resolved to.
`viaDeploymentId`	string \| null	The deployment the request arrived through, `null` when the model was addressed directly.

Polling this endpoint alone is enough to follow a document end to end: extractionStatus covers extraction, and validationExpected plus validationStatus tell you whether to keep waiting for validation or that none is coming.

Rule results

validations holds the model's cross-field, row and document rules, evaluated against the document's current values when you make the request. Field and column rules are not here — their results sit on the cells, on Document Data and Get File Data. For one list covering every scope, read validations.rules[] on Document Data.

Because these results are computed for the request, they reflect corrections submitted since extraction and any edit to the rules themselves, while the stored results on the data endpoints reflect the last evaluation.

Group	Entry
`crossField`	`{ruleId, ruleName, passed, fieldA, fieldB, operator, expected, actual}`, where `fieldA` and `fieldB` are `{columnId, value}` for the two compared fields
`row`	`{ruleId, ruleName, rowResults}`, with one `{rowId, passed, computedValue, targetValue, expected, actual}` per row of the rule's table
`document`	`{ruleId, ruleName, passed, aggregationFunction, computedValue, targetValue, operator, expected, actual}`

ruleName is the rule's name, or the same generated label the other endpoints use when the rule has none — for example Net Amount vs Gross Amount or SUM(Amount). Validations covers how those labels are built.

Get Document Pages

GET/api/v1/deployments/{deploymentId}/documents/{fileId}/pages

Returns page-level access URLs for a document. For PDFs, returns a single PDF URL. For images, returns the image URL directly.

Parameters

Name	Type	In	Required	Description
`deploymentId`	string	path	Yes	The model or deployment ID
`fileId`	string	path	Yes	The file ID

Request

const res = await fetch(
"https://app.limai.io/api/v1/deployments/dep_abc123/documents/file_001/pages",
{
  headers: {
    Authorization: "Bearer YOUR_API_TOKEN",
  },
}
);
const pages = await res.json();

Response

Response200

{
  "fileId": "file_001",
  "fileName": "invoice.pdf",
  "type": "pdf",
  "pageCount": 3,
  "pdfUrl": "https://s3.amazonaws.com/...presigned-url...",
  "pages": []
}

Get Document Content

GET/api/v1/deployments/{deploymentId}/documents/{fileId}/content

Returns the document content including markdown text, PDF URL, and source file URL.

Parameters

Name	Type	In	Required	Description
`deploymentId`	string	path	Yes	The model or deployment ID
`fileId`	string	path	Yes	The file ID

Request

const res = await fetch(
"https://app.limai.io/api/v1/deployments/dep_abc123/documents/file_001/content",
{
  headers: {
    Authorization: "Bearer YOUR_API_TOKEN",
  },
}
);
const content = await res.json();

Response

Response200

{
  "fileId": "file_001",
  "fileName": "invoice.pdf",
  "markdownContent": "# Invoice\\n\\nInvoice Number: INV-2025-001\\n...",
  "pdfUrl": "https://s3.amazonaws.com/...pdf-url...",
  "sourceFileUrl": "https://s3.amazonaws.com/...source-url...",
  "hasMarkdown": true,
  "hasPdf": true,
  "originalExtension": "pdf"
}

Document Detail

On this page