Problems with PowerAutomate/Postman and Stirling PDF - OCR-API Endpoint

[INNMJO]

Hello,

i have a Power Automate Flow which calls OCR-API Endpoint by using classic HTTP-Request. After hours of research i am not getting it to work. As you see in the image, this is my API-Call and the body ist defined in a previous "Compose"-Action. You can see the content below. The outputs('Dateiinhalt_abrufen') is a result of SharePoint Action so it is the file as base64. The problem is not limited to PowerAutomate HTTP-Action but also Postman, when i want to Send the file as base64. Any idea how i can solve this so i can call OCR-API via PowerAutomate Flow?

--boundary123
Content-Disposition: form-data; name="fileInput"; filename="Test.pdf"
Content-Type: application/pdf

@{base64ToBinary(outputs('Dateiinhalt_abrufen')?['body/$content'])}
--boundary123
Content-Disposition: form-data; name="languages"

deu
--boundary123
Content-Disposition: form-data; name="ocrType"

skip-text
--boundary123
Content-Disposition: form-data; name="ocrRenderType"

hocr
--boundary123--

[dosubot[bot]]

Hi @INNMJO! I'm Dosu and I’m helping the Stirling-PDF team.

The Stirling PDF OCR-API endpoint expects the PDF file to be sent as binary data in a multipart/form-data request, not as a base64 string. If you send the file as base64 (even if you use base64ToBinary in PowerAutomate), the API will reject it or return a 400 Bad Request because it does not decode base64 input—it expects a real file upload under the fileInput field with Content-Type: application/pdf docs.

PowerAutomate:
After retrieving your file from SharePoint (which gives you base64), you need to convert it to binary and attach it as a file in the HTTP request. Use the "Compose" action with base64ToBinary(outputs('Dateiinhalt_abrufen')?['body/$content']), then in your HTTP action, set the body type to "form-data" and use the output as a file attachment, not as a string.

Postman:
In Postman, use the "form-data" body type, add a key named fileInput, and use the "File" type to select your PDF file directly. Do not paste base64 content—attach the actual file.

Other requirements:
Include the X-API-KEY header with your API key. Make sure all other fields (languages, ocrType, ocrRenderType) are included as regular form fields.

If you still get a 400 error, double-check that the file is being sent as binary and not as base64 text. You can test the API using Swagger UI at /swagger-ui/index.html on your Stirling-PDF instance to see the expected request format docs.

Let me know if you need a step-by-step for PowerAutomate or have error details!

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}    

[INNMJO]

So it is impossible to do this in PowerAutomate as you cannot change the HTTP action to body=form-data and use a real file as attachment? It has to be possible anyhow...

[dosubot[bot]]

You're correct that PowerAutomate's standard HTTP action can't natively send multipart/form-data with a real binary file attachment the way tools like Postman or curl can. The Stirling PDF API requires the file to be sent as a true binary file in a multipart/form-data request, not as base64 or as a string in the body, and there are no documented workarounds or built-in PowerAutomate features that change this limitation docs.

If you need to automate this, common workarounds are: (1) use a custom connector in PowerAutomate that supports file-type parameters (this requires OpenAPI/Swagger definition and sometimes extra configuration), (2) call an intermediate service (like an Azure Function or Logic App) that receives the file from PowerAutomate and forwards it as multipart/form-data to Stirling PDF, or (3) use a different automation platform that supports file uploads natively.

Let me know if you want more details on any of these approaches!

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}