Sending a PDF to OpenAI /responses endpoint causes 500 error

Question

Sending a PDF to OpenAI /responses endpoint causes 500 error

Scott Smith 25

Hi

I'm trying to use the /responses endpoint of the OpenAI service to analyze a PDF file but it is responding with an Internal Error. I am using the Python openai library (openai-1.76.0). The last request I did (request ID e415f048-be0e-4bd7-9504-81b51a156558) was like this:

response = client.responses.create(

model=deployment,

input=[

{

"role": "user",

"content": [

{

"type": "input_file",

"filename": "page_001.pdf",

"file_data": f"data:application/pdf;base64,{base64_pdf}"

},

{ "type": "input_text", "text": "What kind of document is in this PDF?" }

]

}

]

)

with the response:

{

"error":{

"message":"The server had an error processing your request. Sorry about that! You can retry your request, or contact us through our help center at oai-assistants@microsoft.com if you keep seeing this error. (Please include the request ID e415f048-be0e-4bd7-9504-81b51a156558 in your email.)",

"type":"server_error",

"param":"None",

"code":"None"

}

My OpenAI service is in the UK South region and I'm using gpt-4o and targetting version 2025-03-01-preview.

I've tried other requests including just asking a question and supplying an image which I ask about and they do not cause the error. I only get it when uploading a PDF.

thanks

Scott Smith 25 Reputation points

2025-04-28T13:32:26.4566667+00:00

Thanks for the info, Pavankumar.

Do you know if support for PDFs will be added in the future? I was working with the OpenAI documentation found here https://platform.openai.com/docs/guides/pdf-files?api-mode=responses so just curious if it will eventually mirror this.

thanks
Pavankumar Purilla 11,655 Reputation points Moderator

2025-04-29T07:53:47.3833333+00:00

Hi Scott Smith,
Sorry for inconvenience. At the moment, Microsoft has not announced a release date for stable support of PDFs in Azure OpenAI Service. However, you can keep an eye on the Azure OpenAI Service release notes for updates. Any new features, including PDF support, will be announced there as they become available.
Thank you!

Dieter Devlaminck 15

As of March 2025 this seems to be documented here: https://learn.microsoft.com/en-us/azure/ai-foundry/openai/how-to/responses?tabs=python-secure#file-input

Doing a curl last week this was working well, but today we are also getting the status 500 error.


{
    "model": "deployment",
    "input": [
        {
            "role": "user",
            "content": [
                {
                    "type": "input_file",
                    "filename": "document.pdf",
                    "file_data": "data:application/pdf;base64,{base64-content}"
                },
                {
                    "type": "input_text",
                    "text": "Extract..."
                }
            ]
        }
    ],
    "temperature": 0
}

Without the input_file argument a response comes back

Answer accepted by question author

0 additional answers

Your answer

Scott Smith 25 Reputation points

2025-04-28T13:32:26.4566667+00:00

Thanks for the info, Pavankumar.

Do you know if support for PDFs will be added in the future? I was working with the OpenAI documentation found here https://platform.openai.com/docs/guides/pdf-files?api-mode=responses so just curious if it will eventually mirror this.

thanks
Pavankumar Purilla 11,655 Reputation points Moderator

2025-04-29T07:53:47.3833333+00:00

Hi Scott Smith,
Sorry for inconvenience. At the moment, Microsoft has not announced a release date for stable support of PDFs in Azure OpenAI Service. However, you can keep an eye on the Azure OpenAI Service release notes for updates. Any new features, including PDF support, will be announced there as they become available.
Thank you!
Dieter Devlaminck 15 Reputation points

2025-07-28T15:55:40.7266667+00:00

As of March 2025 this seems to be documented here: https://learn.microsoft.com/en-us/azure/ai-foundry/openai/how-to/responses?tabs=python-secure#file-input

Doing a curl last week this was working well, but today we are also getting the status 500 error.

{ "model": "deployment", "input": [ { "role": "user", "content": [ { "type": "input_file", "filename": "document.pdf", "file_data": "data:application/pdf;base64,{base64-content}" }, { "type": "input_text", "text": "Extract..." } ] } ], "temperature": 0 }

Without the input_file argument a response comes back

Answer 1

Hi Scott Smith,

The reason you're getting a 500 Internal Server Error when sending a PDF to the /responses endpoint is because the Azure OpenAI service does not currently support PDFs as a valid input_file type. The /responses endpoint only accepts certain file types, mainly images like PNG and JPEG, but not PDFs. That’s why sending an image works, but sending a PDF causes the server to crash. To work around this, you should first extract the text content from the PDF using a Python library such as PyMuPDF, PyPDF2, or pdfplumber. Once you have the extracted text, you can send it to the /responses endpoint as input_text instead of input_file. This way, the model can read and analyze the contents of the PDF properly. In short, PDFs must be converted into plain text before sending, because direct PDF uploads are not supported in the /responses API.

I hope this information helps. Thank you!