Which OCR solution should be used to make texts on JIRA-images accessible?

[Besse_ Kees]

LS,

Our JIRA tickets often include text as well as screenshots or images, which themselves frequently contain text. Currently, I can now only search text within the JIRA database.

Could you advise on an OCR solution that would allow us to extract text from these screenshots/images? The goal is to enable queries that cover both the regular ticket text and the text extracted from images.

Thank you in advance for your support.

Best regards,
Kees Besse

[Besse_ Kees]

Additional remark: I noticed there is an app "AI snapshot for JIRA": That will not solve my problem. I want to search through the total JIRA-database and not having to change every individually screenshot. That would take ages.

Regards,

Kees

[Kārlis Rozenbergs]

Hi Kees, I actually built a tool for this called Attachment Architect, but there’s a catch regarding your bulk indexing requirement.

I decided to run the OCR (Tesseract) 100% in the user's browser. This keeps the data strictly private and allowed me to keep the "Runs on Atlassian" badge since no data ever leaves Jira.

The trade-off is that it’s a manual action - a user has to open the preview and trigger the scan to push the text into a comment. Because the processing happens client-side, I can’t really run a background task to index your entire historical database automatically without using an external server, which is exactly what I wanted to avoid for security reasons. It is also not possible to run Tesseract server-side because or Forge limitations and usage costs.

It’s a good fit for making new/active tickets searchable, but honestly, it won't solve your "index everything from the past" problem unless you're willing to do it one by one.