Announcements
{
I am creating a prompt that has multiple questions that I am concatenating with a 4 page PDF document (I OCR'd it). This is a legal document that has the usual types of punctuation. I have no control over the formatting of the PDF. I have disabled "encode request body" in the Invoke Web Service Action. When I run the flow I get the error shown. If I remove the Cleaned Document from the prompt, the flow runs without error, So the problem is with the pdf. I am thinking that I need to escape various characters but am not sure how to proceed. (Are there examples of replace for various types of punctuation.)
"error": {"message": "We could not parse the JSON body of your request. (HINT: This likely means you aren't using your HTTP library correctly. The OpenAI API expects a JSON payload, but what was sent was not valid JSON. If you have trouble figuring out how to fix this, please contact us through our help center at help.openai.com.)","type": "invalid_request_error","param": null,"code": null}}
I am the OP. I copied the PDF into a text file and deleted all the line breaks so it is now one long text string. Still getting error
I am the OP. I made some progress by splitting the Extracted Text on a new line which produces a list. Then I joined the list items with a space. Still getting json error. here is a new flow
Can you share a screenshot of the Invoke web service action properties (clean it for any sensitive info first)? Mainly the body part of it. I don't see any JSON formatting in your flow prior to that and want to make sure you applied appropriate formatting in the body.
-------------------------------------------------------------------------If I have answered your question, please mark it as the preferred solution. If you like my response, please give it a Thumbs Up.
I also provide paid consultancy and development services using Power Automate. If you're interested, DM me and we can discuss it.
Hi Agnius. I was able to solve my problem.
My problem was that my pdf extraction contained carriage returns.
I solved this by using the actions Split Text (on new line) which produces a list
And then join text (space) to regenerate the document without carriage returns.
I then had to crop my text for test purposes to get under the davinci token limit.
This then worked.
Below is my final flow. For test purposes I cropped my input to get under the davinci token limit.
Under review
Thank you for your reply! To ensure a great experience for everyone, your content is awaiting approval by our Community Managers. Please check back later.
Congratulations to our community stars!
Expanding mentorship, skilling, and AI innovation
These are the community rock stars!
Stay up to date on forum activity by subscribing.
David_MA 227 Super User 2026 Season 2
11manish 221 Super User 2026 Season 2
Mohsin Ali 132
