Hello everyone,
I’m developing a web application in Bubble.io as part of my final year project in Social Work, and I’ve encountered a technical problem that I can’t seem to solve, despite trying multiple approaches.
The application involves a professional entering a social assessment in free text. When a button is pressed, this text is sent to OpenAI for analysis and classification into different sections relevant to the field of social work (risk factors, protective factors, emotional state, ADLs/IADLs, economic situation, support network, etc.). Each of these sections should automatically display on its corresponding card within the interface.
The API call works correctly: OpenAI returns the information, and it’s saved in the Bubble database in a text field (raw_text). In the App Data section, I can verify that the AI’s response is there and clearly contains the expected information. The workflow also seems correct: button → API call → record creation → attempt to display the data on the cards.
The problem is that, although the information exists and is saved, I can’t get it to display on the cards. I’ve tried different strategies: returning structured JSON, returning plain text, using regex to extract each section, using custom states, and different ways of linking the container group to the data. In all cases, the cards appear empty in preview or don’t correctly detect the content, even though everything seems to be configured correctly.
My feeling is that I’m doing “the right thing” according to the documentation and the usual examples, but there’s some limitation or detail in Bubble when working with AI responses (long text, JSON, Markdown, etc.) that I’m missing. After quite some time trying to solve it, I no longer know if the problem is with parsing, how Bubble interprets external responses, or some key step I’m overlooking.
I would greatly appreciate any guidance, similar experiences, or suggestions, as this project is crucial for completing my studies and I’m currently feeling quite stuck.
Thank you so much for reading and for any help you can offer.
