AI Document Table Data Extraction

2000 atoms
AI tier


This Utility API uses AI to recognize and extract essential information within tables in documents and images.
This API can analyse visual data to extract tabular information, offering versatility across various applications from optimizing data retrieval in documents to providing context-aware insights in augmented reality experiences.
Note: In order to test the API on API Playground, just click on "Show optional inputs" and enter the Authentication token for the provider before clicking on Send request. The output response structure and the result of the AI utility APIs depend on the service provider and it may vary depending on which service provider is selected.
Endpoint Selector

API Playground


input file: output json

Request example

curl --location --request POST 'https://api.apyhub.com/ai/document/extract/table/file' \
--header 'apy-token: {{token}}' \
--form 'file=@"sample.jpg"'
--form 'requested_service="azure"'
--form 'azure_key="your-azure-key"'
--form 'azure_endpoint="your-azure-endpoint"'
Method: POST
Content Type: multipart/form-data
Request Body
filefileYesthe source document file in any of supported formats (jpeg,jpg, png, tiff, heif, bmp, pdf, docx, xlsx, pptx ,html)
requested_serviceStringyesthe name of service provider for example azure,google
azure_keyStringyes (if azure is selected in requested_service)service key provided by azure
azure_endpointStringyes (if azure is selected in requested_service)the endpoint provided by azure
Sample Response
A successful request returns the extracted table data response in the output parameter specified. If the request fails, the response contains an error code and a message to help determine what went wrong.

HTTP Response Codes

The method may return one of the following HTTP status codes:
Status CodeDescription
200The request was successful.
400Invalid input - the file is corrupt or the supported inputs are not provided.
401Required authentication information is either missing or not valid for the resource.
500If any unexpected error occurs while processing the request.


All requests to ApyHub services need to be authenticated, currently we support tokens or basic authentication mechanisms. You can generate and view your existing credentials from workspace settings (on the left side of the navbar) and go to applications.
Points to note:
  • Credential secrets are generated on the fly and are not stored in plain text, so on generating a credential please save the secrets somewhere safe.
  • Use the apy-token as the header parameter to pass the token.
  • Use the Authorization header to send the basic authentication credentials.

Error codes

"error": {
"code": 105,
"message": "Invalid URL"
To search for a specific error code, enter the code in the search box below. Alternatively, you can click on the button to view a complete list of all error codes.