Dear OCR.space Team,
I am writing to report an issue with my recently purchased PRO OCR API key. The PRO API is consistently failing on images that the Free API processes successfully.
When using the PRO API, every request returns the error message: “E202:Some internal processing error occurred. Please give us some time, we are looking into this!”
To reproduce this, I tested five random sample images. The Free API processed all of them without issue, while the PRO API failed on all five. I have attached my testing script and the sample images (test_ocr.tar) for your review.
After extracting the files in the .tar file (“tar -xvf test_ocr.tar“), the command is as following:
Free API, which can process all files successfully: python ocr_test.py --end_point https://api.ocr.space/parse/image --api_key [my_free_key]
PRO API, which fails on most samples: python ocr_test.py --end_point http://apipro1.ocr.space/parse/image --api_key [my_pro_key]
Could you please investigate this issue and advise on a solution?
Thank you for your time and assistance.
Best regards,
Attachment: the test_ocr.tar is in Google Drive:
Thank you for providing a detailed test case. I tested the images, and they work just fine with the Free and the PRO endpoints, see screenshot:
=> If you test the connection with Postman, do you get the same result?
You probably know it, but just in case: Postman is a free app.
=> If you see the problem even with Postman, a screenshot of the result could be helpful. Because currently I can not recreate your issue.
PS: We are very transparent about our uptime. You can view it at the OCR API Status Dashboard in real time. As expected the PRO uptime is much higher, as the PRO endpoints run on a distributed network of OCR servers, and the servers itself run at a lower load.
Thank you for the prompt reply and your suggestion.
Following your advice, I did some further testing and have an update. You were right that Postman works—all my test cases were successful.
However, the issue persisted when I returned to my Python script. The script, which uses the requests library, continued to fail on approximately 70% of my dataset with the E202 error. This script is a modified version of GitHub - ErikBoesen/ocrspace: A Python API wrapper for the ocr.space Optical Character Recognition API. · GitHub, the 30% random success suggests there is no bugs in this script (100% success in free api with the same script).
Based on the success with Postman, I tried a different approach. I modified my Python script to execute a curl command directly, and this solved the problem completely. All images are now processed successfully.
import os
command = “curl -H apikey:[my_api_key] --form ‘file=@[my_file_path]’ --form ‘language=eng’ --form ‘OCREngine=2’ --form ‘scale=true’ --form ‘isOverlayRequired=true’ --form ‘detectOrientation=true’ https://apipro1.ocr.space/parse/image > [save_root]”
os.system(command)
This suggests there may be a specific issue between the PRO API endpoint and requests made from the Python requests library, as the Free API endpoint does not exhibit this behavior.
My problem is solved with the curl workaround, but I wanted to share these details in case it helps your team investigate a potential backend issue.
Thank you again for your help.
Best regards,
Thanks a lot for this very helpful feedback! We looked into this a bit more and found that the issue happens on the free endpoint as well. But (on all endpoints) “only” for engine1. We added this to our todo list.
