Hi everyone,

Wanted to share a tool we've built that the community might find useful: Pirai AI converts survey documents (Word, PDF, Excel, CSV) directly into LimeSurvey LSS files ready for import.

A bit of background - I've spent about 15 years in market research, so the conversion logic is built on real survey-programming experience, not just generic document parsing. Pirai AI currently supports LimeSurvey 5.x and 6.x (handles the newer normalised schema with the l10ns tables), along with Matrix, Ranking, Multiple Choice, Multiple Numerical, and other question types.

If your survey currently lives in Qualtrics, SurveyMonkey, REDCap, Forsta, or one of 20+ other platforms, you can convert it straight into LimeSurvey too - not just from a Word doc.

Still actively refining the LimeSurvey output based on real-world imports, so if anyone wants to try it on a real survey and let me know what breaks or what's missing, I'd genuinely appreciate it. You can try it here: platform.piraiai.com

Happy to answer any questions about how the conversion works under the hood.

Thanks!
Arun

Looks like a very interesting tool. Will need to test it.

However, what I am missing on your site is clear statements about data privacy and protection. What happens with the surveys that are uploaded? Often the data is confidential, so I think it would be important to highlight what happens with the data, what are you doing with the data, what is the AI doing with the data? Is it used for training? Is it stored, is it private, is it used for other things?

These would be my main worries before uploading a survey document to your platform.

Hi Holch,

Thanks for bringing this up. You're completely right, data privacy is huge for researchers and we took it into account from day one when building this since our background is in enterprise market research.

To answer your questions, the AI only handles structural extraction to map things to the LimeSurvey schema. It doesn't do any content profiling.

We use secure enterprise API endpoints, meaning all data is completely excluded from model training or fine-tuning. It stays entirely private to your session.

For storage, we use a strict data minimization approach. Data is only held temporarily so you can grab your LSS file.Users can request data deletion at any time, but for company accounts, you can actually manage your own automated data retention policy right in the system settings. Everything is encrypted in transit and at rest too.

We have all of these details spelled out on our main website's privacy page under section (piraiai.com/privacy)1.4 if you want to take a look.

If you get a chance to test it out with a dummy survey on the platform, I'd love to hear how the import goes for you!

Cheers,
Arun

Great, this is what I was looking for. Maybe you can put a link to the privacy statement in the footer of the plattform as well. I think this is very helpful. I will have a thorough read through the privacy statement, but based on a first scan, this looks pretty good already.

Thanks Holch,The Privacy Policy link has been added to the platform footer.We’ll keep monitoring how things behave across the latest versions.Thanks again.

Thanks to everyone who has tested it and shared their feedback.

I just tried this with a fairly lengthy Word questionnaire and was amazed at the level of accuracy in generating an lss file for importing to Limesurvey. Very impressive website too.
I had a couple of suggestions for options when generating the lss:
1) Could it set the 'Mandatory' setting to true for all questions?
2) Could it generate each individual question in it's own group?

My surveys always consist of a group and then one question in that group.

I will definitely be doing some more testing as this is a great time saver.

Thanks,
Paul

Hi Paul, thank you so much for trying Pirai AI and for taking the time to share such detailed feedback. I really appreciate you testing it with a lengthy questionnaire and sharing your experience.

The current V1 version supports 9 question types and is focused on accurately converting survey documents into LimeSurvey format. Regarding groups, the converter already identifies the survey structure/blocks and creates separate question groups during import. When importing the LSS file, selecting “Keep the survey group from the imported file” will preserve the groups generated by Pirai AI.

In the upcoming V2, we are expanding support for more advanced survey features, including required/mandatory settings, randomization, “Other specify” options, and additional LimeSurvey configurations.

For the one-question-per-group workflow, we’ll look at adding this as an option during generation, as we understand some teams prefer structuring their LimeSurvey projects this way. It is already possible to achieve this using the Table Edit view, where you can export the converted survey as Excel and bulk update the blocks. Alternatively, we could explore building a dedicated custom extension that automatically creates one question group per question, regardless of the original survey structure.

Thanks again for your thoughtful suggestions and kind words. Feedback like this is extremely valuable and helps us prioritise the next set of features. I hope you continue testing Pirai AI, and I’d love to hear any other ideas or improvements you think would make it more useful.

Best Regards,
Arun