Delete Indexed Document – API Reference

Overview

PurposeFinds all Azure AI Search index chunks belonging to a given document_id and deletes them in a single batch operation.

MethodPOST

AuthHeader — api-key: ankitg.in

Content-TypeNo body required (all params are headers)

Endpoint

POST https://labs.lowcademy.com/apis/delete-indexed-document.php

Request Headers

Header	Required	Type	Description
`api-key`	Required	String	Lab API key. Hardcoded value: `ankitg.in`
`azureAISearchEndpoint`	Required	String (URL)	Base URL of your Azure AI Search service, e.g. `https://my-search.search.windows.net`
`azureAISearchKey`	Required	String	Admin or query key for your Azure AI Search service
`indexName`	Required	String	Name of the Azure AI Search index to target
`document_id`	Required	Integer	Numeric ID of the document whose chunks should be deleted

How It Works

Step	Action
1	Validates all required headers. Returns `400` if any are missing or `document_id` is non-numeric.
2	Calls Azure AI Search Search API with filter `document_id eq {value}` to retrieve up to 1,000 matching chunk IDs.
3	If no chunks are found, returns success with `deleted_chunks: 0`.
4	Calls Azure AI Search Index API with `@search.action: "delete"` for each chunk ID in a single batch request.
5	Returns a JSON response with `success`, `message`, and `deleted_chunks` count.

Sample Requests

cURL

# Delete all indexed chunks for document_id 42
curl -X POST https://labs.lowcademy.com/apis/delete-indexed-document.php \
  -H "api-key: ankitg.in" \
  -H "azureAISearchEndpoint: https://my-search.search.windows.net" \
  -H "azureAISearchKey: YOUR_SEARCH_ADMIN_KEY" \
  -H "indexName: my-documents-index" \
  -H "document_id: 42"

JavaScript (fetch)

const response = await fetch('https://labs.lowcademy.com/apis/delete-indexed-document.php', {
  method: 'POST',
  headers: {
    'api-key':                  'ankitg.in',
    'azureAISearchEndpoint':  'https://my-search.search.windows.net',
    'azureAISearchKey':        'YOUR_SEARCH_ADMIN_KEY',
    'indexName':               'my-documents-index',
    'document_id':             '42',
  }
});
const result = await response.json();
console.log(result);

OutSystems (REST API call)

// Method: POST
// URL: https://labs.lowcademy.com/apis/delete-indexed-document.php
// Headers (map from OutSystems input parameters):
api-key                 → "ankitg.in"               // hardcoded
azureAISearchEndpoint   → AzureAISearchEndpoint     // Text input
azureAISearchKey        → AzureAISearchKey           // Text input
indexName               → IndexName                  // Text input
document_id             → DocumentId                 // Integer input → ToString

Response Schema

Success — chunks deleted (HTTP 200)

{
  "success":        true,
  "message":        "Successfully deleted 7 chunk(s) for document_id 42 from index 'my-documents-index'.",
  "deleted_chunks": 7
}

Success — nothing to delete (HTTP 200)

{
  "success":        true,
  "message":        "No indexed chunks found for document_id 42. Nothing was deleted.",
  "deleted_chunks": 0
}

Error (HTTP 4xx / 5xx)

{
  "success": false,
  "message": "Missing required header: indexName."
}

Response Codes

Code	Meaning
`200`	Success — chunks deleted, or no chunks found (both are non-error states)
`400`	Missing or invalid header parameter
`401`	Invalid or missing `api-key`
`405`	Wrong HTTP method (use POST)
`502`	Azure AI Search returned an error (search or delete call failed)

Notes

Batch limit: Up to 1,000 chunks per document are deleted in a single Azure AI Search batch call. Documents with more than 1,000 chunks would require multiple calls (not expected in typical usage — a 1,200-char chunk of a 300-page PDF produces ~750 chunks at most).

document_id filter: Azure AI Search uses OData filter syntax for Int64 fields — the value must be unquoted: document_id eq 42 (not document_id eq '42'). This API handles that automatically.

Idempotent: Calling this API twice for the same document_id is safe. The second call returns deleted_chunks: 0 with success.